hipblas  v4.0.5.0
prototype interfaces compatible with ROCm platform and HiP
Classes | Macros | Typedefs | Enumerations | Functions
hipblas.h File Reference
#include "hipblas-export.h"
#include "hipblas-version.h"
#include <hip/hip_runtime_api.h>
#include <stdint.h>

Go to the source code of this file.

Classes

struct  hipblasBfloat16
 Struct to represent a 16 bit Brain floating-point number. More...
 
struct  hipblasComplex
 Struct to represent a complex number with single precision real and imaginary parts. More...
 
struct  hipblasDoubleComplex
 Struct to represent a complex number with double precision real and imaginary parts. More...
 

Macros

#define HIPBLAS_CLANG_STATIC
 

Typedefs

typedef void * hipblasHandle_t
 hipblasHanlde_t is a void pointer, to store the library context (either rocBLAS or cuBLAS) More...
 
typedef uint16_t hipblasHalf
 To specify the datatype to be unsigned short. More...
 
typedef int8_t hipblasInt8
 To specify the datatype to be signed char. More...
 
typedef int64_t hipblasStride
 Stride between matrices or vectors in strided_batched functions. More...
 

Enumerations

enum  hipblasStatus_t {
  HIPBLAS_STATUS_SUCCESS = 0,
  HIPBLAS_STATUS_NOT_INITIALIZED = 1,
  HIPBLAS_STATUS_ALLOC_FAILED = 2,
  HIPBLAS_STATUS_INVALID_VALUE = 3,
  HIPBLAS_STATUS_MAPPING_ERROR = 4,
  HIPBLAS_STATUS_EXECUTION_FAILED = 5,
  HIPBLAS_STATUS_INTERNAL_ERROR = 6,
  HIPBLAS_STATUS_NOT_SUPPORTED = 7,
  HIPBLAS_STATUS_ARCH_MISMATCH = 8,
  HIPBLAS_STATUS_HANDLE_IS_NULLPTR = 9,
  HIPBLAS_STATUS_INVALID_ENUM = 10,
  HIPBLAS_STATUS_UNKNOWN = 11
}
 hipblas status codes definition More...
 
enum  hipblasOperation_t {
  HIPBLAS_OP_N = 111,
  HIPBLAS_OP_T = 112,
  HIPBLAS_OP_C = 113
}
 Used to specify whether the matrix is to be transposed or not. More...
 
enum  hipblasPointerMode_t {
  HIPBLAS_POINTER_MODE_HOST,
  HIPBLAS_POINTER_MODE_DEVICE
}
 Indicates if scalar pointers are on host or device. This is used for scalars alpha and beta and for scalar function return values. More...
 
enum  hipblasFillMode_t {
  HIPBLAS_FILL_MODE_UPPER = 121,
  HIPBLAS_FILL_MODE_LOWER = 122,
  HIPBLAS_FILL_MODE_FULL = 123
}
 Used by the Hermitian, symmetric and triangular matrix routines to specify whether the upper or lower triangle is being referenced. More...
 
enum  hipblasDiagType_t {
  HIPBLAS_DIAG_NON_UNIT = 131,
  HIPBLAS_DIAG_UNIT = 132
}
 It is used by the triangular matrix routines to specify whether the matrix is unit triangular. More...
 
enum  hipblasSideMode_t {
  HIPBLAS_SIDE_LEFT = 141,
  HIPBLAS_SIDE_RIGHT = 142,
  HIPBLAS_SIDE_BOTH = 143
}
 Indicates the side matrix A is located relative to matrix B during multiplication. More...
 
enum  hipblasDatatype_t {
  HIPBLAS_R_16F = 150,
  HIPBLAS_R_32F = 151,
  HIPBLAS_R_64F = 152,
  HIPBLAS_C_16F = 153,
  HIPBLAS_C_32F = 154,
  HIPBLAS_C_64F = 155,
  HIPBLAS_R_8I = 160,
  HIPBLAS_R_8U = 161,
  HIPBLAS_R_32I = 162,
  HIPBLAS_R_32U = 163,
  HIPBLAS_C_8I = 164,
  HIPBLAS_C_8U = 165,
  HIPBLAS_C_32I = 166,
  HIPBLAS_C_32U = 167,
  HIPBLAS_R_16B = 168,
  HIPBLAS_C_16B = 169,
  HIPBLAS_DATATYPE_INVALID = 255
}
 Indicates the precision width of data stored in a blas type. More...
 
enum  hipblasGemmAlgo_t { HIPBLAS_GEMM_DEFAULT = 160 }
 Indicates if layer is active with bitmask. More...
 
enum  hipblasAtomicsMode_t {
  HIPBLAS_ATOMICS_NOT_ALLOWED = 0,
  HIPBLAS_ATOMICS_ALLOWED = 1
}
 Indicates if atomics operations are allowed. Not allowing atomic operations may generally improve determinism and repeatability of results at a cost of performance. More...
 

Functions

hipblasStatus_t hipblasCreate (hipblasHandle_t *handle)
 Create hipblas handle. More...
 
hipblasStatus_t hipblasDestroy (hipblasHandle_t handle)
 Destroys the library context created using hipblasCreate() More...
 
hipblasStatus_t hipblasSetStream (hipblasHandle_t handle, hipStream_t streamId)
 Set stream for handle. More...
 
hipblasStatus_t hipblasGetStream (hipblasHandle_t handle, hipStream_t *streamId)
 Get stream[0] for handle. More...
 
hipblasStatus_t hipblasSetPointerMode (hipblasHandle_t handle, hipblasPointerMode_t mode)
 Set hipblas pointer mode. More...
 
hipblasStatus_t hipblasGetPointerMode (hipblasHandle_t handle, hipblasPointerMode_t *mode)
 Get hipblas pointer mode. More...
 
hipblasStatus_t hipblasSetVector (int n, int elemSize, const void *x, int incx, void *y, int incy)
 copy vector from host to device More...
 
hipblasStatus_t hipblasGetVector (int n, int elemSize, const void *x, int incx, void *y, int incy)
 copy vector from device to host More...
 
hipblasStatus_t hipblasSetMatrix (int rows, int cols, int elemSize, const void *AP, int lda, void *BP, int ldb)
 copy matrix from host to device More...
 
hipblasStatus_t hipblasGetMatrix (int rows, int cols, int elemSize, const void *AP, int lda, void *BP, int ldb)
 copy matrix from device to host More...
 
hipblasStatus_t hipblasSetVectorAsync (int n, int elemSize, const void *x, int incx, void *y, int incy, hipStream_t stream)
 asynchronously copy vector from host to device More...
 
hipblasStatus_t hipblasGetVectorAsync (int n, int elemSize, const void *x, int incx, void *y, int incy, hipStream_t stream)
 asynchronously copy vector from device to host More...
 
hipblasStatus_t hipblasSetMatrixAsync (int rows, int cols, int elemSize, const void *AP, int lda, void *BP, int ldb, hipStream_t stream)
 asynchronously copy matrix from host to device More...
 
hipblasStatus_t hipblasGetMatrixAsync (int rows, int cols, int elemSize, const void *AP, int lda, void *BP, int ldb, hipStream_t stream)
 asynchronously copy matrix from device to host More...
 
hipblasStatus_t hipblasSetAtomicsMode (hipblasHandle_t handle, hipblasAtomicsMode_t atomics_mode)
 Set hipblasSetAtomicsMode. More...
 
hipblasStatus_t hipblasGetAtomicsMode (hipblasHandle_t handle, hipblasAtomicsMode_t *atomics_mode)
 Get hipblasSetAtomicsMode. More...
 
hipblasStatus_t hipblasStrmmOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const float *alpha, const float *AP, int lda, const float *BP, int ldb, float *CP, int ldc)
 
hipblasStatus_t hipblasDtrmmOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const double *alpha, const double *AP, int lda, const double *BP, int ldb, double *CP, int ldc)
 
hipblasStatus_t hipblasCtrmmOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *BP, int ldb, hipblasComplex *CP, int ldc)
 
hipblasStatus_t hipblasZtrmmOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *BP, int ldb, hipblasDoubleComplex *CP, int ldc)
 
hipblasStatus_t hipblasStrmmBatchedOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const float *alpha, const float *const AP[], int lda, const float *const BP[], int ldb, float *const CP[], int ldc, int batchCount)
 
hipblasStatus_t hipblasDtrmmBatchedOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const double *alpha, const double *const AP[], int lda, const double *const BP[], int ldb, double *const CP[], int ldc, int batchCount)
 
hipblasStatus_t hipblasCtrmmBatchedOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const BP[], int ldb, hipblasComplex *const CP[], int ldc, int batchCount)
 
hipblasStatus_t hipblasZtrmmBatchedOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const BP[], int ldb, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 
hipblasStatus_t hipblasStrmmStridedBatchedOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const float *alpha, const float *AP, int lda, hipblasStride strideA, const float *BP, int ldb, hipblasStride strideB, float *CP, int ldc, hipblasStride strideC, int batchCount)
 
hipblasStatus_t hipblasDtrmmStridedBatchedOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const double *alpha, const double *AP, int lda, hipblasStride strideA, const double *BP, int ldb, hipblasStride strideB, double *CP, int ldc, hipblasStride strideC, int batchCount)
 
hipblasStatus_t hipblasCtrmmStridedBatchedOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *BP, int ldb, hipblasStride strideB, hipblasComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 
hipblasStatus_t hipblasZtrmmStridedBatchedOutofplace (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *BP, int ldb, hipblasStride strideB, hipblasDoubleComplex *BC, int ldc, hipblasStride strideC, int batchCount)
 
hipblasStatus_t hipblasGemmEx (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const void *alpha, const void *A, hipblasDatatype_t aType, int lda, const void *B, hipblasDatatype_t bType, int ldb, const void *beta, void *C, hipblasDatatype_t cType, int ldc, hipblasDatatype_t computeType, hipblasGemmAlgo_t algo)
 BLAS EX API. More...
 
hipblasStatus_t hipblasGemmBatchedEx (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const void *alpha, const void *A[], hipblasDatatype_t aType, int lda, const void *B[], hipblasDatatype_t bType, int ldb, const void *beta, void *C[], hipblasDatatype_t cType, int ldc, int batchCount, hipblasDatatype_t computeType, hipblasGemmAlgo_t algo)
 BLAS EX API. More...
 
hipblasStatus_t hipblasGemmStridedBatchedEx (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const void *alpha, const void *A, hipblasDatatype_t aType, int lda, hipblasStride strideA, const void *B, hipblasDatatype_t bType, int ldb, hipblasStride strideB, const void *beta, void *C, hipblasDatatype_t cType, int ldc, hipblasStride strideC, int batchCount, hipblasDatatype_t computeType, hipblasGemmAlgo_t algo)
 BLAS EX API. More...
 
hipblasStatus_t hipblasTrsmEx (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const void *alpha, void *A, int lda, void *B, int ldb, const void *invA, int invAsize, hipblasDatatype_t computeType)
 
hipblasStatus_t hipblasTrsmBatchedEx (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const void *alpha, void *A, int lda, void *B, int ldb, int batchCount, const void *invA, int invAsize, hipblasDatatype_t computeType)
 
hipblasStatus_t hipblasTrsmStridedBatchedEx (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const void *alpha, void *A, int lda, hipblasStride strideA, void *B, int ldb, hipblasStride strideB, int batchCount, const void *invA, int invAsize, hipblasStride strideInvA, hipblasDatatype_t computeType)
 
hipblasStatus_t hipblasAxpyEx (hipblasHandle_t handle, int n, const void *alpha, hipblasDatatype_t alphaType, const void *x, hipblasDatatype_t xType, int incx, void *y, hipblasDatatype_t yType, int incy, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasAxpyBatchedEx (hipblasHandle_t handle, int n, const void *alpha, hipblasDatatype_t alphaType, const void *x, hipblasDatatype_t xType, int incx, void *y, hipblasDatatype_t yType, int incy, int batchCount, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasAxpyStridedBatchedEx (hipblasHandle_t handle, int n, const void *alpha, hipblasDatatype_t alphaType, const void *x, hipblasDatatype_t xType, int incx, hipblasStride stridex, void *y, hipblasDatatype_t yType, int incy, hipblasStride stridey, int batchCount, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasNrm2Ex (hipblasHandle_t handle, int n, const void *x, hipblasDatatype_t xType, int incx, void *result, hipblasDatatype_t resultType, hipblasDatatype_t executionType)
 BLAS_EX API. More...
 
hipblasStatus_t hipblasNrm2BatchedEx (hipblasHandle_t handle, int n, const void *x, hipblasDatatype_t xType, int incx, int batchCount, void *result, hipblasDatatype_t resultType, hipblasDatatype_t executionType)
 BLAS_EX API. More...
 
hipblasStatus_t hipblasNrm2StridedBatchedEx (hipblasHandle_t handle, int n, const void *x, hipblasDatatype_t xType, int incx, hipblasStride stridex, int batchCount, void *result, hipblasDatatype_t resultType, hipblasDatatype_t executionType)
 BLAS_EX API. More...
 
hipblasStatus_t hipblasRotEx (hipblasHandle_t handle, int n, void *x, hipblasDatatype_t xType, int incx, void *y, hipblasDatatype_t yType, int incy, const void *c, const void *s, hipblasDatatype_t csType, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasRotBatchedEx (hipblasHandle_t handle, int n, void *x, hipblasDatatype_t xType, int incx, void *y, hipblasDatatype_t yType, int incy, const void *c, const void *s, hipblasDatatype_t csType, int batchCount, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasRotStridedBatchedEx (hipblasHandle_t handle, int n, void *x, hipblasDatatype_t xType, int incx, hipblasStride stridex, void *y, hipblasDatatype_t yType, int incy, hipblasStride stridey, const void *c, const void *s, hipblasDatatype_t csType, int batchCount, hipblasDatatype_t executionType)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasScalEx (hipblasHandle_t handle, int n, const void *alpha, hipblasDatatype_t alphaType, void *x, hipblasDatatype_t xType, int incx, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasScalBatchedEx (hipblasHandle_t handle, int n, const void *alpha, hipblasDatatype_t alphaType, void *x, hipblasDatatype_t xType, int incx, int batchCount, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasScalStridedBatchedEx (hipblasHandle_t handle, int n, const void *alpha, hipblasDatatype_t alphaType, void *x, hipblasDatatype_t xType, int incx, hipblasStride stridex, int batchCount, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
const char * hipblasStatusToString (hipblasStatus_t status)
 
hipblasStatus_t hipblasIsamax (hipblasHandle_t handle, int n, const float *x, int incx, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIdamax (hipblasHandle_t handle, int n, const double *x, int incx, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIcamax (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIzamax (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIsamaxBatched (hipblasHandle_t handle, int n, const float *const x[], int incx, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIdamaxBatched (hipblasHandle_t handle, int n, const double *const x[], int incx, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIcamaxBatched (hipblasHandle_t handle, int n, const hipblasComplex *const x[], int incx, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIzamaxBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *const x[], int incx, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIsamaxStridedBatched (hipblasHandle_t handle, int n, const float *x, int incx, hipblasStride stridex, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIdamaxStridedBatched (hipblasHandle_t handle, int n, const double *x, int incx, hipblasStride stridex, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIcamaxStridedBatched (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, hipblasStride stridex, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIzamaxStridedBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIsamin (hipblasHandle_t handle, int n, const float *x, int incx, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIdamin (hipblasHandle_t handle, int n, const double *x, int incx, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIcamin (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIzamin (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIsaminBatched (hipblasHandle_t handle, int n, const float *const x[], int incx, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIdaminBatched (hipblasHandle_t handle, int n, const double *const x[], int incx, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIcaminBatched (hipblasHandle_t handle, int n, const hipblasComplex *const x[], int incx, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIzaminBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *const x[], int incx, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIsaminStridedBatched (hipblasHandle_t handle, int n, const float *x, int incx, hipblasStride stridex, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIdaminStridedBatched (hipblasHandle_t handle, int n, const double *x, int incx, hipblasStride stridex, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIcaminStridedBatched (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, hipblasStride stridex, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasIzaminStridedBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount, int *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSasum (hipblasHandle_t handle, int n, const float *x, int incx, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDasum (hipblasHandle_t handle, int n, const double *x, int incx, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasScasum (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDzasum (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSasumBatched (hipblasHandle_t handle, int n, const float *const x[], int incx, int batchCount, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDasumBatched (hipblasHandle_t handle, int n, const double *const x[], int incx, int batchCount, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasScasumBatched (hipblasHandle_t handle, int n, const hipblasComplex *const x[], int incx, int batchCount, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDzasumBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *const x[], int incx, int batchCount, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSasumStridedBatched (hipblasHandle_t handle, int n, const float *x, int incx, hipblasStride stridex, int batchCount, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDasumStridedBatched (hipblasHandle_t handle, int n, const double *x, int incx, hipblasStride stridex, int batchCount, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasScasumStridedBatched (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, hipblasStride stridex, int batchCount, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDzasumStridedBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasHaxpy (hipblasHandle_t handle, int n, const hipblasHalf *alpha, const hipblasHalf *x, int incx, hipblasHalf *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSaxpy (hipblasHandle_t handle, int n, const float *alpha, const float *x, int incx, float *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDaxpy (hipblasHandle_t handle, int n, const double *alpha, const double *x, int incx, double *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCaxpy (hipblasHandle_t handle, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, hipblasComplex *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZaxpy (hipblasHandle_t handle, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, hipblasDoubleComplex *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasHaxpyBatched (hipblasHandle_t handle, int n, const hipblasHalf *alpha, const hipblasHalf *const x[], int incx, hipblasHalf *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSaxpyBatched (hipblasHandle_t handle, int n, const float *alpha, const float *const x[], int incx, float *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDaxpyBatched (hipblasHandle_t handle, int n, const double *alpha, const double *const x[], int incx, double *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCaxpyBatched (hipblasHandle_t handle, int n, const hipblasComplex *alpha, const hipblasComplex *const x[], int incx, hipblasComplex *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZaxpyBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const x[], int incx, hipblasDoubleComplex *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasHaxpyStridedBatched (hipblasHandle_t handle, int n, const hipblasHalf *alpha, const hipblasHalf *x, int incx, hipblasStride stridex, hipblasHalf *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSaxpyStridedBatched (hipblasHandle_t handle, int n, const float *alpha, const float *x, int incx, hipblasStride stridex, float *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDaxpyStridedBatched (hipblasHandle_t handle, int n, const double *alpha, const double *x, int incx, hipblasStride stridex, double *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCaxpyStridedBatched (hipblasHandle_t handle, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, hipblasStride stridex, hipblasComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZaxpyStridedBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, hipblasDoubleComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasScopy (hipblasHandle_t handle, int n, const float *x, int incx, float *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDcopy (hipblasHandle_t handle, int n, const double *x, int incx, double *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCcopy (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, hipblasComplex *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZcopy (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, hipblasDoubleComplex *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasScopyBatched (hipblasHandle_t handle, int n, const float *const x[], int incx, float *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDcopyBatched (hipblasHandle_t handle, int n, const double *const x[], int incx, double *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCcopyBatched (hipblasHandle_t handle, int n, const hipblasComplex *const x[], int incx, hipblasComplex *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZcopyBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *const x[], int incx, hipblasDoubleComplex *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasScopyStridedBatched (hipblasHandle_t handle, int n, const float *x, int incx, hipblasStride stridex, float *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDcopyStridedBatched (hipblasHandle_t handle, int n, const double *x, int incx, hipblasStride stridex, double *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCcopyStridedBatched (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, hipblasStride stridex, hipblasComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZcopyStridedBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, hipblasDoubleComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasHdot (hipblasHandle_t handle, int n, const hipblasHalf *x, int incx, const hipblasHalf *y, int incy, hipblasHalf *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasBfdot (hipblasHandle_t handle, int n, const hipblasBfloat16 *x, int incx, const hipblasBfloat16 *y, int incy, hipblasBfloat16 *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSdot (hipblasHandle_t handle, int n, const float *x, int incx, const float *y, int incy, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDdot (hipblasHandle_t handle, int n, const double *x, int incx, const double *y, int incy, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCdotc (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, const hipblasComplex *y, int incy, hipblasComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCdotu (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, const hipblasComplex *y, int incy, hipblasComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdotc (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *y, int incy, hipblasDoubleComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdotu (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *y, int incy, hipblasDoubleComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasHdotBatched (hipblasHandle_t handle, int n, const hipblasHalf *const x[], int incx, const hipblasHalf *const y[], int incy, int batchCount, hipblasHalf *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasBfdotBatched (hipblasHandle_t handle, int n, const hipblasBfloat16 *const x[], int incx, const hipblasBfloat16 *const y[], int incy, int batchCount, hipblasBfloat16 *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSdotBatched (hipblasHandle_t handle, int n, const float *const x[], int incx, const float *const y[], int incy, int batchCount, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDdotBatched (hipblasHandle_t handle, int n, const double *const x[], int incx, const double *const y[], int incy, int batchCount, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCdotcBatched (hipblasHandle_t handle, int n, const hipblasComplex *const x[], int incx, const hipblasComplex *const y[], int incy, int batchCount, hipblasComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCdotuBatched (hipblasHandle_t handle, int n, const hipblasComplex *const x[], int incx, const hipblasComplex *const y[], int incy, int batchCount, hipblasComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdotcBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *const y[], int incy, int batchCount, hipblasDoubleComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdotuBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *const y[], int incy, int batchCount, hipblasDoubleComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasHdotStridedBatched (hipblasHandle_t handle, int n, const hipblasHalf *x, int incx, hipblasStride stridex, const hipblasHalf *y, int incy, hipblasStride stridey, int batchCount, hipblasHalf *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasBfdotStridedBatched (hipblasHandle_t handle, int n, const hipblasBfloat16 *x, int incx, hipblasStride stridex, const hipblasBfloat16 *y, int incy, hipblasStride stridey, int batchCount, hipblasBfloat16 *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSdotStridedBatched (hipblasHandle_t handle, int n, const float *x, int incx, hipblasStride stridex, const float *y, int incy, hipblasStride stridey, int batchCount, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDdotStridedBatched (hipblasHandle_t handle, int n, const double *x, int incx, hipblasStride stridex, const double *y, int incy, hipblasStride stridey, int batchCount, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCdotcStridedBatched (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *y, int incy, hipblasStride stridey, int batchCount, hipblasComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCdotuStridedBatched (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *y, int incy, hipblasStride stridey, int batchCount, hipblasComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdotcStridedBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *y, int incy, hipblasStride stridey, int batchCount, hipblasDoubleComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdotuStridedBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *y, int incy, hipblasStride stridey, int batchCount, hipblasDoubleComplex *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSnrm2 (hipblasHandle_t handle, int n, const float *x, int incx, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDnrm2 (hipblasHandle_t handle, int n, const double *x, int incx, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasScnrm2 (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDznrm2 (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSnrm2Batched (hipblasHandle_t handle, int n, const float *const x[], int incx, int batchCount, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDnrm2Batched (hipblasHandle_t handle, int n, const double *const x[], int incx, int batchCount, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasScnrm2Batched (hipblasHandle_t handle, int n, const hipblasComplex *const x[], int incx, int batchCount, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDznrm2Batched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *const x[], int incx, int batchCount, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSnrm2StridedBatched (hipblasHandle_t handle, int n, const float *x, int incx, hipblasStride stridex, int batchCount, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDnrm2StridedBatched (hipblasHandle_t handle, int n, const double *x, int incx, hipblasStride stridex, int batchCount, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasScnrm2StridedBatched (hipblasHandle_t handle, int n, const hipblasComplex *x, int incx, hipblasStride stridex, int batchCount, float *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDznrm2StridedBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount, double *result)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrot (hipblasHandle_t handle, int n, float *x, int incx, float *y, int incy, const float *c, const float *s)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrot (hipblasHandle_t handle, int n, double *x, int incx, double *y, int incy, const double *c, const double *s)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCrot (hipblasHandle_t handle, int n, hipblasComplex *x, int incx, hipblasComplex *y, int incy, const float *c, const hipblasComplex *s)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCsrot (hipblasHandle_t handle, int n, hipblasComplex *x, int incx, hipblasComplex *y, int incy, const float *c, const float *s)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZrot (hipblasHandle_t handle, int n, hipblasDoubleComplex *x, int incx, hipblasDoubleComplex *y, int incy, const double *c, const hipblasDoubleComplex *s)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdrot (hipblasHandle_t handle, int n, hipblasDoubleComplex *x, int incx, hipblasDoubleComplex *y, int incy, const double *c, const double *s)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrotBatched (hipblasHandle_t handle, int n, float *const x[], int incx, float *const y[], int incy, const float *c, const float *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrotBatched (hipblasHandle_t handle, int n, double *const x[], int incx, double *const y[], int incy, const double *c, const double *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCrotBatched (hipblasHandle_t handle, int n, hipblasComplex *const x[], int incx, hipblasComplex *const y[], int incy, const float *c, const hipblasComplex *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCsrotBatched (hipblasHandle_t handle, int n, hipblasComplex *const x[], int incx, hipblasComplex *const y[], int incy, const float *c, const float *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZrotBatched (hipblasHandle_t handle, int n, hipblasDoubleComplex *const x[], int incx, hipblasDoubleComplex *const y[], int incy, const double *c, const hipblasDoubleComplex *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdrotBatched (hipblasHandle_t handle, int n, hipblasDoubleComplex *const x[], int incx, hipblasDoubleComplex *const y[], int incy, const double *c, const double *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrotStridedBatched (hipblasHandle_t handle, int n, float *x, int incx, hipblasStride stridex, float *y, int incy, hipblasStride stridey, const float *c, const float *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrotStridedBatched (hipblasHandle_t handle, int n, double *x, int incx, hipblasStride stridex, double *y, int incy, hipblasStride stridey, const double *c, const double *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCrotStridedBatched (hipblasHandle_t handle, int n, hipblasComplex *x, int incx, hipblasStride stridex, hipblasComplex *y, int incy, hipblasStride stridey, const float *c, const hipblasComplex *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCsrotStridedBatched (hipblasHandle_t handle, int n, hipblasComplex *x, int incx, hipblasStride stridex, hipblasComplex *y, int incy, hipblasStride stridey, const float *c, const float *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZrotStridedBatched (hipblasHandle_t handle, int n, hipblasDoubleComplex *x, int incx, hipblasStride stridex, hipblasDoubleComplex *y, int incy, hipblasStride stridey, const double *c, const hipblasDoubleComplex *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdrotStridedBatched (hipblasHandle_t handle, int n, hipblasDoubleComplex *x, int incx, hipblasStride stridex, hipblasDoubleComplex *y, int incy, hipblasStride stridey, const double *c, const double *s, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrotg (hipblasHandle_t handle, float *a, float *b, float *c, float *s)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrotg (hipblasHandle_t handle, double *a, double *b, double *c, double *s)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCrotg (hipblasHandle_t handle, hipblasComplex *a, hipblasComplex *b, float *c, hipblasComplex *s)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZrotg (hipblasHandle_t handle, hipblasDoubleComplex *a, hipblasDoubleComplex *b, double *c, hipblasDoubleComplex *s)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrotgBatched (hipblasHandle_t handle, float *const a[], float *const b[], float *const c[], float *const s[], int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrotgBatched (hipblasHandle_t handle, double *const a[], double *const b[], double *const c[], double *const s[], int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCrotgBatched (hipblasHandle_t handle, hipblasComplex *const a[], hipblasComplex *const b[], float *const c[], hipblasComplex *const s[], int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZrotgBatched (hipblasHandle_t handle, hipblasDoubleComplex *const a[], hipblasDoubleComplex *const b[], double *const c[], hipblasDoubleComplex *const s[], int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrotgStridedBatched (hipblasHandle_t handle, float *a, hipblasStride stridea, float *b, hipblasStride strideb, float *c, hipblasStride stridec, float *s, hipblasStride strides, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrotgStridedBatched (hipblasHandle_t handle, double *a, hipblasStride stridea, double *b, hipblasStride strideb, double *c, hipblasStride stridec, double *s, hipblasStride strides, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCrotgStridedBatched (hipblasHandle_t handle, hipblasComplex *a, hipblasStride stridea, hipblasComplex *b, hipblasStride strideb, float *c, hipblasStride stridec, hipblasComplex *s, hipblasStride strides, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZrotgStridedBatched (hipblasHandle_t handle, hipblasDoubleComplex *a, hipblasStride stridea, hipblasDoubleComplex *b, hipblasStride strideb, double *c, hipblasStride stridec, hipblasDoubleComplex *s, hipblasStride strides, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrotm (hipblasHandle_t handle, int n, float *x, int incx, float *y, int incy, const float *param)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrotm (hipblasHandle_t handle, int n, double *x, int incx, double *y, int incy, const double *param)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrotmBatched (hipblasHandle_t handle, int n, float *const x[], int incx, float *const y[], int incy, const float *const param[], int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrotmBatched (hipblasHandle_t handle, int n, double *const x[], int incx, double *const y[], int incy, const double *const param[], int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrotmStridedBatched (hipblasHandle_t handle, int n, float *x, int incx, hipblasStride stridex, float *y, int incy, hipblasStride stridey, const float *param, hipblasStride strideParam, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrotmStridedBatched (hipblasHandle_t handle, int n, double *x, int incx, hipblasStride stridex, double *y, int incy, hipblasStride stridey, const double *param, hipblasStride strideParam, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrotmg (hipblasHandle_t handle, float *d1, float *d2, float *x1, const float *y1, float *param)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrotmg (hipblasHandle_t handle, double *d1, double *d2, double *x1, const double *y1, double *param)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrotmgBatched (hipblasHandle_t handle, float *const d1[], float *const d2[], float *const x1[], const float *const y1[], float *const param[], int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrotmgBatched (hipblasHandle_t handle, double *const d1[], double *const d2[], double *const x1[], const double *const y1[], double *const param[], int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSrotmgStridedBatched (hipblasHandle_t handle, float *d1, hipblasStride strided1, float *d2, hipblasStride strided2, float *x1, hipblasStride stridex1, const float *y1, hipblasStride stridey1, float *param, hipblasStride strideParam, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDrotmgStridedBatched (hipblasHandle_t handle, double *d1, hipblasStride strided1, double *d2, hipblasStride strided2, double *x1, hipblasStride stridex1, const double *y1, hipblasStride stridey1, double *param, hipblasStride strideParam, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSscal (hipblasHandle_t handle, int n, const float *alpha, float *x, int incx)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDscal (hipblasHandle_t handle, int n, const double *alpha, double *x, int incx)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCscal (hipblasHandle_t handle, int n, const hipblasComplex *alpha, hipblasComplex *x, int incx)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCsscal (hipblasHandle_t handle, int n, const float *alpha, hipblasComplex *x, int incx)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZscal (hipblasHandle_t handle, int n, const hipblasDoubleComplex *alpha, hipblasDoubleComplex *x, int incx)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdscal (hipblasHandle_t handle, int n, const double *alpha, hipblasDoubleComplex *x, int incx)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSscalBatched (hipblasHandle_t handle, int n, const float *alpha, float *const x[], int incx, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDscalBatched (hipblasHandle_t handle, int n, const double *alpha, double *const x[], int incx, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCscalBatched (hipblasHandle_t handle, int n, const hipblasComplex *alpha, hipblasComplex *const x[], int incx, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZscalBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *alpha, hipblasDoubleComplex *const x[], int incx, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCsscalBatched (hipblasHandle_t handle, int n, const float *alpha, hipblasComplex *const x[], int incx, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdscalBatched (hipblasHandle_t handle, int n, const double *alpha, hipblasDoubleComplex *const x[], int incx, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSscalStridedBatched (hipblasHandle_t handle, int n, const float *alpha, float *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDscalStridedBatched (hipblasHandle_t handle, int n, const double *alpha, double *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCscalStridedBatched (hipblasHandle_t handle, int n, const hipblasComplex *alpha, hipblasComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZscalStridedBatched (hipblasHandle_t handle, int n, const hipblasDoubleComplex *alpha, hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCsscalStridedBatched (hipblasHandle_t handle, int n, const float *alpha, hipblasComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZdscalStridedBatched (hipblasHandle_t handle, int n, const double *alpha, hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSswap (hipblasHandle_t handle, int n, float *x, int incx, float *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDswap (hipblasHandle_t handle, int n, double *x, int incx, double *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCswap (hipblasHandle_t handle, int n, hipblasComplex *x, int incx, hipblasComplex *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZswap (hipblasHandle_t handle, int n, hipblasDoubleComplex *x, int incx, hipblasDoubleComplex *y, int incy)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSswapBatched (hipblasHandle_t handle, int n, float *const x[], int incx, float *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDswapBatched (hipblasHandle_t handle, int n, double *const x[], int incx, double *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCswapBatched (hipblasHandle_t handle, int n, hipblasComplex *const x[], int incx, hipblasComplex *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZswapBatched (hipblasHandle_t handle, int n, hipblasDoubleComplex *const x[], int incx, hipblasDoubleComplex *const y[], int incy, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSswapStridedBatched (hipblasHandle_t handle, int n, float *x, int incx, hipblasStride stridex, float *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasDswapStridedBatched (hipblasHandle_t handle, int n, double *x, int incx, hipblasStride stridex, double *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasCswapStridedBatched (hipblasHandle_t handle, int n, hipblasComplex *x, int incx, hipblasStride stridex, hipblasComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasZswapStridedBatched (hipblasHandle_t handle, int n, hipblasDoubleComplex *x, int incx, hipblasStride stridex, hipblasDoubleComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 1 API. More...
 
hipblasStatus_t hipblasSgbmv (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const float *alpha, const float *AP, int lda, const float *x, int incx, const float *beta, float *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDgbmv (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const double *alpha, const double *AP, int lda, const double *x, int incx, const double *beta, double *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgbmv (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *x, int incx, const hipblasComplex *beta, hipblasComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgbmv (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSgbmvBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const float *alpha, const float *const AP[], int lda, const float *const x[], int incx, const float *beta, float *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDgbmvBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const double *alpha, const double *const AP[], int lda, const double *const x[], int incx, const double *beta, double *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgbmvBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const x[], int incx, const hipblasComplex *beta, hipblasComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgbmvBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSgbmvStridedBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const float *alpha, const float *AP, int lda, hipblasStride strideA, const float *x, int incx, hipblasStride stridex, const float *beta, float *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDgbmvStridedBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const double *alpha, const double *AP, int lda, hipblasStride strideA, const double *x, int incx, hipblasStride stridex, const double *beta, double *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgbmvStridedBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *beta, hipblasComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgbmvStridedBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, int kl, int ku, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSgemv (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, const float *alpha, const float *AP, int lda, const float *x, int incx, const float *beta, float *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDgemv (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, const double *alpha, const double *AP, int lda, const double *x, int incx, const double *beta, double *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgemv (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *x, int incx, const hipblasComplex *beta, hipblasComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgemv (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSgemvBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, const float *alpha, const float *const AP[], int lda, const float *const x[], int incx, const float *beta, float *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDgemvBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, const double *alpha, const double *const AP[], int lda, const double *const x[], int incx, const double *beta, double *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgemvBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const x[], int incx, const hipblasComplex *beta, hipblasComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgemvBatched (hipblasHandle_t handle, hipblasOperation_t trans, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSgemvStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, int m, int n, const float *alpha, const float *AP, int lda, hipblasStride strideA, const float *x, int incx, hipblasStride stridex, const float *beta, float *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDgemvStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, int m, int n, const double *alpha, const double *AP, int lda, hipblasStride strideA, const double *x, int incx, hipblasStride stridex, const double *beta, double *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgemvStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *beta, hipblasComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgemvStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSger (hipblasHandle_t handle, int m, int n, const float *alpha, const float *x, int incx, const float *y, int incy, float *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDger (hipblasHandle_t handle, int m, int n, const double *alpha, const double *x, int incx, const double *y, int incy, double *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgeru (hipblasHandle_t handle, int m, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, const hipblasComplex *y, int incy, hipblasComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgerc (hipblasHandle_t handle, int m, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, const hipblasComplex *y, int incy, hipblasComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgeru (hipblasHandle_t handle, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *y, int incy, hipblasDoubleComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgerc (hipblasHandle_t handle, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *y, int incy, hipblasDoubleComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSgerBatched (hipblasHandle_t handle, int m, int n, const float *alpha, const float *const x[], int incx, const float *const y[], int incy, float *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDgerBatched (hipblasHandle_t handle, int m, int n, const double *alpha, const double *const x[], int incx, const double *const y[], int incy, double *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgeruBatched (hipblasHandle_t handle, int m, int n, const hipblasComplex *alpha, const hipblasComplex *const x[], int incx, const hipblasComplex *const y[], int incy, hipblasComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgercBatched (hipblasHandle_t handle, int m, int n, const hipblasComplex *alpha, const hipblasComplex *const x[], int incx, const hipblasComplex *const y[], int incy, hipblasComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgeruBatched (hipblasHandle_t handle, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *const y[], int incy, hipblasDoubleComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgercBatched (hipblasHandle_t handle, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *const y[], int incy, hipblasDoubleComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSgerStridedBatched (hipblasHandle_t handle, int m, int n, const float *alpha, const float *x, int incx, hipblasStride stridex, const float *y, int incy, hipblasStride stridey, float *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDgerStridedBatched (hipblasHandle_t handle, int m, int n, const double *alpha, const double *x, int incx, hipblasStride stridex, const double *y, int incy, hipblasStride stridey, double *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgeruStridedBatched (hipblasHandle_t handle, int m, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *y, int incy, hipblasStride stridey, hipblasComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCgercStridedBatched (hipblasHandle_t handle, int m, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *y, int incy, hipblasStride stridey, hipblasComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgeruStridedBatched (hipblasHandle_t handle, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *y, int incy, hipblasStride stridey, hipblasDoubleComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZgercStridedBatched (hipblasHandle_t handle, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *y, int incy, hipblasStride stridey, hipblasDoubleComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChbmv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *x, int incx, const hipblasComplex *beta, hipblasComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhbmv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChbmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const x[], int incx, const hipblasComplex *beta, hipblasComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhbmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChbmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *beta, hipblasComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhbmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChemv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *x, int incx, const hipblasComplex *beta, hipblasComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhemv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChemvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const x[], int incx, const hipblasComplex *beta, hipblasComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhemvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChemvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *beta, hipblasComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhemvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCher (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const hipblasComplex *x, int incx, hipblasComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZher (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const hipblasDoubleComplex *x, int incx, hipblasDoubleComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCherBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const hipblasComplex *const x[], int incx, hipblasComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZherBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const hipblasDoubleComplex *const x[], int incx, hipblasDoubleComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCherStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const hipblasComplex *x, int incx, hipblasStride stridex, hipblasComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZherStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, hipblasDoubleComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCher2 (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, const hipblasComplex *y, int incy, hipblasComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZher2 (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *y, int incy, hipblasDoubleComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCher2Batched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *const x[], int incx, const hipblasComplex *const y[], int incy, hipblasComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZher2Batched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *const y[], int incy, hipblasDoubleComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCher2StridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *y, int incy, hipblasStride stridey, hipblasComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZher2StridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *y, int incy, hipblasStride stridey, hipblasDoubleComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChpmv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *AP, const hipblasComplex *x, int incx, const hipblasComplex *beta, hipblasComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhpmv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChpmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *const AP[], const hipblasComplex *const x[], int incx, const hipblasComplex *beta, hipblasComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhpmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChpmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *AP, hipblasStride strideA, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *beta, hipblasComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhpmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, hipblasStride strideA, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChpr (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const hipblasComplex *x, int incx, hipblasComplex *AP)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhpr (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const hipblasDoubleComplex *x, int incx, hipblasDoubleComplex *AP)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChprBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const hipblasComplex *const x[], int incx, hipblasComplex *const AP[], int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhprBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const hipblasDoubleComplex *const x[], int incx, hipblasDoubleComplex *const AP[], int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChprStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const hipblasComplex *x, int incx, hipblasStride stridex, hipblasComplex *AP, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhprStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, hipblasDoubleComplex *AP, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChpr2 (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, const hipblasComplex *y, int incy, hipblasComplex *AP)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhpr2 (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *y, int incy, hipblasDoubleComplex *AP)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChpr2Batched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *const x[], int incx, const hipblasComplex *const y[], int incy, hipblasComplex *const AP[], int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhpr2Batched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *const y[], int incy, hipblasDoubleComplex *const AP[], int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasChpr2StridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *y, int incy, hipblasStride stridey, hipblasComplex *AP, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZhpr2StridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *y, int incy, hipblasStride stridey, hipblasDoubleComplex *AP, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsbmv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const float *alpha, const float *AP, int lda, const float *x, int incx, const float *beta, float *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsbmv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const double *alpha, const double *AP, int lda, const double *x, int incx, const double *beta, double *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsbmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const float *alpha, const float *const AP[], int lda, const float *const x[], int incx, const float *beta, float *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsbmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const double *alpha, const double *const AP[], int lda, const double *const x[], int incx, const double *beta, double *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsbmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const float *alpha, const float *AP, int lda, hipblasStride strideA, const float *x, int incx, hipblasStride stridex, const float *beta, float *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsbmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, int k, const double *alpha, const double *AP, int lda, hipblasStride strideA, const double *x, int incx, hipblasStride stridex, const double *beta, double *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSspmv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *AP, const float *x, int incx, const float *beta, float *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDspmv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *AP, const double *x, int incx, const double *beta, double *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSspmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *const AP[], const float *const x[], int incx, const float *beta, float *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDspmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *const AP[], const double *const x[], int incx, const double *beta, double *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSspmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *AP, hipblasStride strideA, const float *x, int incx, hipblasStride stridex, const float *beta, float *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDspmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *AP, hipblasStride strideA, const double *x, int incx, hipblasStride stridex, const double *beta, double *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSspr (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *x, int incx, float *AP)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDspr (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *x, int incx, double *AP)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCspr (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, hipblasComplex *AP)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZspr (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, hipblasDoubleComplex *AP)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsprBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *const x[], int incx, float *const AP[], int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsprBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *const x[], int incx, double *const AP[], int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCsprBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *const x[], int incx, hipblasComplex *const AP[], int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZsprBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const x[], int incx, hipblasDoubleComplex *const AP[], int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsprStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *x, int incx, hipblasStride stridex, float *AP, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsprStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *x, int incx, hipblasStride stridex, double *AP, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCsprStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, hipblasStride stridex, hipblasComplex *AP, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZsprStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, hipblasDoubleComplex *AP, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSspr2 (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *x, int incx, const float *y, int incy, float *AP)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDspr2 (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *x, int incx, const double *y, int incy, double *AP)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSspr2Batched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *const x[], int incx, const float *const y[], int incy, float *const AP[], int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDspr2Batched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *const x[], int incx, const double *const y[], int incy, double *const AP[], int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSspr2StridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *x, int incx, hipblasStride stridex, const float *y, int incy, hipblasStride stridey, float *AP, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDspr2StridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *x, int incx, hipblasStride stridex, const double *y, int incy, hipblasStride stridey, double *AP, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsymv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *AP, int lda, const float *x, int incx, const float *beta, float *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsymv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *AP, int lda, const double *x, int incx, const double *beta, double *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCsymv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *x, int incx, const hipblasComplex *beta, hipblasComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZsymv (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsymvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *const AP[], int lda, const float *const x[], int incx, const float *beta, float *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsymvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *const AP[], int lda, const double *const x[], int incx, const double *beta, double *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCsymvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const x[], int incx, const hipblasComplex *beta, hipblasComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZsymvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const y[], int incy, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsymvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *AP, int lda, hipblasStride strideA, const float *x, int incx, hipblasStride stridex, const float *beta, float *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsymvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *AP, int lda, hipblasStride strideA, const double *x, int incx, hipblasStride stridex, const double *beta, double *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCsymvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *beta, hipblasComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZsymvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *beta, hipblasDoubleComplex *y, int incy, hipblasStride stridey, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsyr (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *x, int incx, float *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsyr (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *x, int incx, double *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCsyr (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, hipblasComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZsyr (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, hipblasDoubleComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsyrBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *const x[], int incx, float *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsyrBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *const x[], int incx, double *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCsyrBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *const x[], int incx, hipblasComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZsyrBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const x[], int incx, hipblasDoubleComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsyrStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *x, int incx, hipblasStride stridex, float *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsyrStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *x, int incx, hipblasStride stridex, double *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCsyrStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, hipblasStride stridex, hipblasComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZsyrStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, hipblasDoubleComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsyr2 (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *x, int incx, const float *y, int incy, float *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsyr2 (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *x, int incx, const double *y, int incy, double *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCsyr2 (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, const hipblasComplex *y, int incy, hipblasComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZsyr2 (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, const hipblasDoubleComplex *y, int incy, hipblasDoubleComplex *AP, int lda)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsyr2Batched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *const x[], int incx, const float *const y[], int incy, float *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsyr2Batched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *const x[], int incx, const double *const y[], int incy, double *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCsyr2Batched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *const x[], int incx, const hipblasComplex *const y[], int incy, hipblasComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZsyr2Batched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const x[], int incx, const hipblasDoubleComplex *const y[], int incy, hipblasDoubleComplex *const AP[], int lda, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasSsyr2StridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const float *alpha, const float *x, int incx, hipblasStride stridex, const float *y, int incy, hipblasStride stridey, float *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDsyr2StridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const double *alpha, const double *x, int incx, hipblasStride stridex, const double *y, int incy, hipblasStride stridey, double *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCsyr2StridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasComplex *alpha, const hipblasComplex *x, int incx, hipblasStride stridex, const hipblasComplex *y, int incy, hipblasStride stridey, hipblasComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZsyr2StridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, const hipblasDoubleComplex *y, int incy, hipblasStride stridey, hipblasDoubleComplex *AP, int lda, hipblasStride strideA, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStbmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const float *AP, int lda, float *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtbmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const double *AP, int lda, double *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtbmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const hipblasComplex *AP, int lda, hipblasComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtbmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const hipblasDoubleComplex *AP, int lda, hipblasDoubleComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStbmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const float *const AP[], int lda, float *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtbmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const double *const AP[], int lda, double *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtbmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const hipblasComplex *const AP[], int lda, hipblasComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtbmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const hipblasDoubleComplex *const AP[], int lda, hipblasDoubleComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStbmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const float *AP, int lda, hipblasStride strideA, float *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtbmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const double *AP, int lda, hipblasStride strideA, double *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtbmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const hipblasComplex *AP, int lda, hipblasStride strideA, hipblasComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtbmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int k, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStbsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const float *AP, int lda, float *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtbsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const double *AP, int lda, double *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtbsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const hipblasComplex *AP, int lda, hipblasComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtbsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const hipblasDoubleComplex *AP, int lda, hipblasDoubleComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStbsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const float *const AP[], int lda, float *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtbsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const double *const AP[], int lda, double *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtbsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const hipblasComplex *const AP[], int lda, hipblasComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtbsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const hipblasDoubleComplex *const AP[], int lda, hipblasDoubleComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStbsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const float *AP, int lda, hipblasStride strideA, float *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtbsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const double *AP, int lda, hipblasStride strideA, double *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtbsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const hipblasComplex *AP, int lda, hipblasStride strideA, hipblasComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtbsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int n, int k, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStpmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *AP, float *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtpmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *AP, double *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtpmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *AP, hipblasComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtpmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *AP, hipblasDoubleComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStpmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *const AP[], float *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtpmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *const AP[], double *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtpmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *const AP[], hipblasComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtpmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *const AP[], hipblasDoubleComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStpmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *AP, hipblasStride strideA, float *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtpmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *AP, hipblasStride strideA, double *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtpmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *AP, hipblasStride strideA, hipblasComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtpmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *AP, hipblasStride strideA, hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStpsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *AP, float *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtpsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *AP, double *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtpsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *AP, hipblasComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtpsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *AP, hipblasDoubleComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStpsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *const AP[], float *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtpsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *const AP[], double *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtpsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *const AP[], hipblasComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtpsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *const AP[], hipblasDoubleComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStpsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *AP, hipblasStride strideA, float *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtpsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *AP, hipblasStride strideA, double *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtpsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *AP, hipblasStride strideA, hipblasComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtpsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *AP, hipblasStride strideA, hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStrmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *AP, int lda, float *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtrmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *AP, int lda, double *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtrmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *AP, int lda, hipblasComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtrmv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *AP, int lda, hipblasDoubleComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStrmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *const AP[], int lda, float *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtrmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *const AP[], int lda, double *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtrmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *const AP[], int lda, hipblasComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtrmvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *const AP[], int lda, hipblasDoubleComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStrmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *AP, int lda, hipblasStride strideA, float *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtrmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *AP, int lda, hipblasStride strideA, double *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtrmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *AP, int lda, hipblasStride strideA, hipblasComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtrmvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStrsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *AP, int lda, float *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtrsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *AP, int lda, double *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtrsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *AP, int lda, hipblasComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtrsv (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *AP, int lda, hipblasDoubleComplex *x, int incx)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStrsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *const AP[], int lda, float *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtrsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *const AP[], int lda, double *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtrsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *const AP[], int lda, hipblasComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtrsvBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *const AP[], int lda, hipblasDoubleComplex *const x[], int incx, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasStrsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const float *AP, int lda, hipblasStride strideA, float *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasDtrsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const double *AP, int lda, hipblasStride strideA, double *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasCtrsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasComplex *AP, int lda, hipblasStride strideA, hipblasComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasZtrsvStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, hipblasDoubleComplex *x, int incx, hipblasStride stridex, int batchCount)
 BLAS Level 2 API. More...
 
hipblasStatus_t hipblasHgemm (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const hipblasHalf *alpha, const hipblasHalf *AP, int lda, const hipblasHalf *BP, int ldb, const hipblasHalf *beta, hipblasHalf *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSgemm (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const float *alpha, const float *AP, int lda, const float *BP, int ldb, const float *beta, float *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDgemm (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const double *alpha, const double *AP, int lda, const double *BP, int ldb, const double *beta, double *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCgemm (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *BP, int ldb, const hipblasComplex *beta, hipblasComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZgemm (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *BP, int ldb, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasHgemmBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const hipblasHalf *alpha, const hipblasHalf *const AP[], int lda, const hipblasHalf *const BP[], int ldb, const hipblasHalf *beta, hipblasHalf *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSgemmBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const float *alpha, const float *const AP[], int lda, const float *const BP[], int ldb, const float *beta, float *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDgemmBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const double *alpha, const double *const AP[], int lda, const double *const BP[], int ldb, const double *beta, double *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCgemmBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const BP[], int ldb, const hipblasComplex *beta, hipblasComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZgemmBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const BP[], int ldb, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasHgemmStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const hipblasHalf *alpha, const hipblasHalf *AP, int lda, long long strideA, const hipblasHalf *BP, int ldb, long long strideB, const hipblasHalf *beta, hipblasHalf *CP, int ldc, long long strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSgemmStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const float *alpha, const float *AP, int lda, long long strideA, const float *BP, int ldb, long long strideB, const float *beta, float *CP, int ldc, long long strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDgemmStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const double *alpha, const double *AP, int lda, long long strideA, const double *BP, int ldb, long long strideB, const double *beta, double *CP, int ldc, long long strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCgemmStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, long long strideA, const hipblasComplex *BP, int ldb, long long strideB, const hipblasComplex *beta, hipblasComplex *CP, int ldc, long long strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZgemmStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, long long strideA, const hipblasDoubleComplex *BP, int ldb, long long strideB, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc, long long strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCherk (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const hipblasComplex *AP, int lda, const float *beta, hipblasComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZherk (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const hipblasDoubleComplex *AP, int lda, const double *beta, hipblasDoubleComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCherkBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const hipblasComplex *const AP[], int lda, const float *beta, hipblasComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZherkBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const hipblasDoubleComplex *const AP[], int lda, const double *beta, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCherkStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const float *beta, hipblasComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZherkStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const double *beta, hipblasDoubleComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCherkx (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *BP, int ldb, const float *beta, hipblasComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZherkx (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *BP, int ldb, const double *beta, hipblasDoubleComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCherkxBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const BP[], int ldb, const float *beta, hipblasComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZherkxBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const BP[], int ldb, const double *beta, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCherkxStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *BP, int ldb, hipblasStride strideB, const float *beta, hipblasComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZherkxStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *BP, int ldb, hipblasStride strideB, const double *beta, hipblasDoubleComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCher2k (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *BP, int ldb, const float *beta, hipblasComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZher2k (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *BP, int ldb, const double *beta, hipblasDoubleComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCher2kBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const BP[], int ldb, const float *beta, hipblasComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZher2kBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const BP[], int ldb, const double *beta, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCher2kStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *BP, int ldb, hipblasStride strideB, const float *beta, hipblasComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZher2kStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *BP, int ldb, hipblasStride strideB, const double *beta, hipblasDoubleComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsymm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const float *alpha, const float *AP, int lda, const float *BP, int ldb, const float *beta, float *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsymm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const double *alpha, const double *AP, int lda, const double *BP, int ldb, const double *beta, double *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsymm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *BP, int ldb, const hipblasComplex *beta, hipblasComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsymm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *BP, int ldb, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsymmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const float *alpha, const float *const AP[], int lda, const float *const BP[], int ldb, const float *beta, float *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsymmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const double *alpha, const double *const AP[], int lda, const double *const BP[], int ldb, const double *beta, double *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsymmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const BP[], int ldb, const hipblasComplex *beta, hipblasComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsymmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const BP[], int ldb, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsymmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const float *alpha, const float *AP, int lda, hipblasStride strideA, const float *BP, int ldb, hipblasStride strideB, const float *beta, float *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsymmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const double *alpha, const double *AP, int lda, hipblasStride strideA, const double *BP, int ldb, hipblasStride strideB, const double *beta, double *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsymmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *BP, int ldb, hipblasStride strideB, const hipblasComplex *beta, hipblasComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsymmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *BP, int ldb, hipblasStride strideB, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsyrk (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const float *AP, int lda, const float *beta, float *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsyrk (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const double *AP, int lda, const double *beta, double *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsyrk (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *beta, hipblasComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsyrk (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsyrkBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const float *const AP[], int lda, const float *beta, float *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsyrkBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const double *const AP[], int lda, const double *beta, double *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsyrkBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *beta, hipblasComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsyrkBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsyrkStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const float *AP, int lda, hipblasStride strideA, const float *beta, float *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsyrkStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const double *AP, int lda, hipblasStride strideA, const double *beta, double *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsyrkStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *beta, hipblasComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsyrkStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsyr2k (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const float *AP, int lda, const float *BP, int ldb, const float *beta, float *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsyr2k (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const double *AP, int lda, const double *BP, int ldb, const double *beta, double *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsyr2k (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *BP, int ldb, const hipblasComplex *beta, hipblasComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsyr2k (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *BP, int ldb, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsyr2kBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const float *const AP[], int lda, const float *const BP[], int ldb, const float *beta, float *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsyr2kBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const double *const AP[], int lda, const double *const BP[], int ldb, const double *beta, double *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsyr2kBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const BP[], int ldb, const hipblasComplex *beta, hipblasComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsyr2kBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const BP[], int ldb, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsyr2kStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const float *AP, int lda, hipblasStride strideA, const float *BP, int ldb, hipblasStride strideB, const float *beta, float *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsyr2kStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const double *AP, int lda, hipblasStride strideA, const double *BP, int ldb, hipblasStride strideB, const double *beta, double *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsyr2kStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *BP, int ldb, hipblasStride strideB, const hipblasComplex *beta, hipblasComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsyr2kStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *BP, int ldb, hipblasStride strideB, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsyrkx (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const float *AP, int lda, const float *BP, int ldb, const float *beta, float *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsyrkx (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const double *AP, int lda, const double *BP, int ldb, const double *beta, double *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsyrkx (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *BP, int ldb, const hipblasComplex *beta, hipblasComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsyrkx (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *BP, int ldb, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsyrkxBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const float *const AP[], int lda, const float *const BP[], int ldb, const float *beta, float *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsyrkxBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const double *const AP[], int lda, const double *const BP[], int ldb, const double *beta, double *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsyrkxBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const BP[], int ldb, const hipblasComplex *beta, hipblasComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsyrkxBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const BP[], int ldb, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSsyrkxStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const float *alpha, const float *AP, int lda, hipblasStride strideA, const float *BP, int ldb, hipblasStride strideB, const float *beta, float *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDsyrkxStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const double *alpha, const double *AP, int lda, hipblasStride strideA, const double *BP, int ldb, hipblasStride strideB, const double *beta, double *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCsyrkxStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *BP, int ldb, hipblasStride strideB, const hipblasComplex *beta, hipblasComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZsyrkxStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasOperation_t transA, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *BP, int ldb, hipblasStride strideB, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSgeam (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const float *alpha, const float *AP, int lda, const float *beta, const float *BP, int ldb, float *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDgeam (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const double *alpha, const double *AP, int lda, const double *beta, const double *BP, int ldb, double *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCgeam (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *beta, const hipblasComplex *BP, int ldb, hipblasComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZgeam (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *beta, const hipblasDoubleComplex *BP, int ldb, hipblasDoubleComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSgeamBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const float *alpha, const float *const AP[], int lda, const float *beta, const float *const BP[], int ldb, float *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDgeamBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const double *alpha, const double *const AP[], int lda, const double *beta, const double *const BP[], int ldb, double *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCgeamBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *beta, const hipblasComplex *const BP[], int ldb, hipblasComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZgeamBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *beta, const hipblasDoubleComplex *const BP[], int ldb, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSgeamStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const float *alpha, const float *AP, int lda, hipblasStride strideA, const float *beta, const float *BP, int ldb, hipblasStride strideB, float *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDgeamStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const double *alpha, const double *AP, int lda, hipblasStride strideA, const double *beta, const double *BP, int ldb, hipblasStride strideB, double *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCgeamStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *beta, const hipblasComplex *BP, int ldb, hipblasStride strideB, hipblasComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZgeamStridedBatched (hipblasHandle_t handle, hipblasOperation_t transA, hipblasOperation_t transB, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *beta, const hipblasDoubleComplex *BP, int ldb, hipblasStride strideB, hipblasDoubleComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasChemm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, const hipblasComplex *BP, int ldb, const hipblasComplex *beta, hipblasComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZhemm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *BP, int ldb, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasChemmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int n, int k, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, const hipblasComplex *const BP[], int ldb, const hipblasComplex *beta, hipblasComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZhemmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const BP[], int ldb, const hipblasDoubleComplex *beta, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasChemmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int n, int k, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *BP, int ldb, hipblasStride strideB, const hipblasComplex *beta, hipblasComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZhemmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, int n, int k, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *BP, int ldb, hipblasStride strideB, const hipblasDoubleComplex *beta, hipblasDoubleComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasStrmm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const float *alpha, const float *AP, int lda, float *BP, int ldb)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDtrmm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const double *alpha, const double *AP, int lda, double *BP, int ldb)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCtrmm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasComplex *BP, int ldb)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZtrmm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasDoubleComplex *BP, int ldb)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasStrmmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const float *alpha, const float *const AP[], int lda, float *const BP[], int ldb, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDtrmmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const double *alpha, const double *const AP[], int lda, double *const BP[], int ldb, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCtrmmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, hipblasComplex *const BP[], int ldb, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZtrmmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, hipblasDoubleComplex *const BP[], int ldb, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasStrmmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const float *alpha, const float *AP, int lda, hipblasStride strideA, float *BP, int ldb, hipblasStride strideB, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDtrmmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const double *alpha, const double *AP, int lda, hipblasStride strideA, double *BP, int ldb, hipblasStride strideB, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCtrmmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, hipblasComplex *BP, int ldb, hipblasStride strideB, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZtrmmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, hipblasDoubleComplex *BP, int ldb, hipblasStride strideB, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasStrsm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const float *alpha, const float *AP, int lda, float *BP, int ldb)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDtrsm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const double *alpha, const double *AP, int lda, double *BP, int ldb)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCtrsm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasComplex *BP, int ldb)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZtrsm (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasDoubleComplex *BP, int ldb)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasStrsmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const float *alpha, const float *const AP[], int lda, float *const BP[], int ldb, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDtrsmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const double *alpha, const double *const AP[], int lda, double *const BP[], int ldb, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCtrsmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasComplex *alpha, const hipblasComplex *const AP[], int lda, hipblasComplex *const BP[], int ldb, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZtrsmBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *const AP[], int lda, hipblasDoubleComplex *const BP[], int ldb, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasStrsmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const float *alpha, const float *AP, int lda, hipblasStride strideA, float *BP, int ldb, hipblasStride strideB, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDtrsmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const double *alpha, const double *AP, int lda, hipblasStride strideA, double *BP, int ldb, hipblasStride strideB, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCtrsmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasComplex *alpha, const hipblasComplex *AP, int lda, hipblasStride strideA, hipblasComplex *BP, int ldb, hipblasStride strideB, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZtrsmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, hipblasFillMode_t uplo, hipblasOperation_t transA, hipblasDiagType_t diag, int m, int n, const hipblasDoubleComplex *alpha, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, hipblasDoubleComplex *BP, int ldb, hipblasStride strideB, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasStrtri (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const float *AP, int lda, float *invA, int ldinvA)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDtrtri (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const double *AP, int lda, double *invA, int ldinvA)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCtrtri (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const hipblasComplex *AP, int lda, hipblasComplex *invA, int ldinvA)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZtrtri (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const hipblasDoubleComplex *AP, int lda, hipblasDoubleComplex *invA, int ldinvA)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasStrtriBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const float *const AP[], int lda, float *invA[], int ldinvA, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDtrtriBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const double *const AP[], int lda, double *invA[], int ldinvA, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCtrtriBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const hipblasComplex *const AP[], int lda, hipblasComplex *invA[], int ldinvA, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZtrtriBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const hipblasDoubleComplex *const AP[], int lda, hipblasDoubleComplex *invA[], int ldinvA, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasStrtriStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const float *AP, int lda, hipblasStride strideA, float *invA, int ldinvA, hipblasStride stride_invA, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDtrtriStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const double *AP, int lda, hipblasStride strideA, double *invA, int ldinvA, hipblasStride stride_invA, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCtrtriStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const hipblasComplex *AP, int lda, hipblasStride strideA, hipblasComplex *invA, int ldinvA, hipblasStride stride_invA, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZtrtriStridedBatched (hipblasHandle_t handle, hipblasFillMode_t uplo, hipblasDiagType_t diag, int n, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, hipblasDoubleComplex *invA, int ldinvA, hipblasStride stride_invA, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSdgmm (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const float *AP, int lda, const float *x, int incx, float *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDdgmm (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const double *AP, int lda, const double *x, int incx, double *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCdgmm (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const hipblasComplex *AP, int lda, const hipblasComplex *x, int incx, hipblasComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZdgmm (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const hipblasDoubleComplex *AP, int lda, const hipblasDoubleComplex *x, int incx, hipblasDoubleComplex *CP, int ldc)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSdgmmBatched (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const float *const AP[], int lda, const float *const x[], int incx, float *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDdgmmBatched (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const double *const AP[], int lda, const double *const x[], int incx, double *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCdgmmBatched (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const hipblasComplex *const AP[], int lda, const hipblasComplex *const x[], int incx, hipblasComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZdgmmBatched (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const hipblasDoubleComplex *const AP[], int lda, const hipblasDoubleComplex *const x[], int incx, hipblasDoubleComplex *const CP[], int ldc, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSdgmmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const float *AP, int lda, hipblasStride strideA, const float *x, int incx, hipblasStride stridex, float *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasDdgmmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const double *AP, int lda, hipblasStride strideA, const double *x, int incx, hipblasStride stridex, double *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasCdgmmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const hipblasComplex *AP, int lda, hipblasStride strideA, const hipblasComplex *x, int incx, hipblasStride stridex, hipblasComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasZdgmmStridedBatched (hipblasHandle_t handle, hipblasSideMode_t side, int m, int n, const hipblasDoubleComplex *AP, int lda, hipblasStride strideA, const hipblasDoubleComplex *x, int incx, hipblasStride stridex, hipblasDoubleComplex *CP, int ldc, hipblasStride strideC, int batchCount)
 BLAS Level 3 API. More...
 
hipblasStatus_t hipblasSgetrf (hipblasHandle_t handle, const int n, float *A, const int lda, int *ipiv, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasDgetrf (hipblasHandle_t handle, const int n, double *A, const int lda, int *ipiv, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasCgetrf (hipblasHandle_t handle, const int n, hipblasComplex *A, const int lda, int *ipiv, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasZgetrf (hipblasHandle_t handle, const int n, hipblasDoubleComplex *A, const int lda, int *ipiv, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasSgetrfBatched (hipblasHandle_t handle, const int n, float *const A[], const int lda, int *ipiv, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasDgetrfBatched (hipblasHandle_t handle, const int n, double *const A[], const int lda, int *ipiv, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasCgetrfBatched (hipblasHandle_t handle, const int n, hipblasComplex *const A[], const int lda, int *ipiv, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasZgetrfBatched (hipblasHandle_t handle, const int n, hipblasDoubleComplex *const A[], const int lda, int *ipiv, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasSgetrfStridedBatched (hipblasHandle_t handle, const int n, float *A, const int lda, const hipblasStride strideA, int *ipiv, const hipblasStride strideP, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasDgetrfStridedBatched (hipblasHandle_t handle, const int n, double *A, const int lda, const hipblasStride strideA, int *ipiv, const hipblasStride strideP, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasCgetrfStridedBatched (hipblasHandle_t handle, const int n, hipblasComplex *A, const int lda, const hipblasStride strideA, int *ipiv, const hipblasStride strideP, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasZgetrfStridedBatched (hipblasHandle_t handle, const int n, hipblasDoubleComplex *A, const int lda, const hipblasStride strideA, int *ipiv, const hipblasStride strideP, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasSgetrs (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, float *A, const int lda, const int *ipiv, float *B, const int ldb, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasDgetrs (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, double *A, const int lda, const int *ipiv, double *B, const int ldb, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasCgetrs (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, hipblasComplex *A, const int lda, const int *ipiv, hipblasComplex *B, const int ldb, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasZgetrs (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, hipblasDoubleComplex *A, const int lda, const int *ipiv, hipblasDoubleComplex *B, const int ldb, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasSgetrsBatched (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, float *const A[], const int lda, const int *ipiv, float *const B[], const int ldb, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasDgetrsBatched (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, double *const A[], const int lda, const int *ipiv, double *const B[], const int ldb, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasCgetrsBatched (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, hipblasComplex *const A[], const int lda, const int *ipiv, hipblasComplex *const B[], const int ldb, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasZgetrsBatched (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, hipblasDoubleComplex *const A[], const int lda, const int *ipiv, hipblasDoubleComplex *const B[], const int ldb, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasSgetrsStridedBatched (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, float *A, const int lda, const hipblasStride strideA, const int *ipiv, const hipblasStride strideP, float *B, const int ldb, const hipblasStride strideB, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasDgetrsStridedBatched (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, double *A, const int lda, const hipblasStride strideA, const int *ipiv, const hipblasStride strideP, double *B, const int ldb, const hipblasStride strideB, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasCgetrsStridedBatched (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, hipblasComplex *A, const int lda, const hipblasStride strideA, const int *ipiv, const hipblasStride strideP, hipblasComplex *B, const int ldb, const hipblasStride strideB, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasZgetrsStridedBatched (hipblasHandle_t handle, const hipblasOperation_t trans, const int n, const int nrhs, hipblasDoubleComplex *A, const int lda, const hipblasStride strideA, const int *ipiv, const hipblasStride strideP, hipblasDoubleComplex *B, const int ldb, const hipblasStride strideB, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasSgetriBatched (hipblasHandle_t handle, const int n, float *const A[], const int lda, int *ipiv, float *const C[], const int ldc, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasDgetriBatched (hipblasHandle_t handle, const int n, double *const A[], const int lda, int *ipiv, double *const C[], const int ldc, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasCgetriBatched (hipblasHandle_t handle, const int n, hipblasComplex *const A[], const int lda, int *ipiv, hipblasComplex *const C[], const int ldc, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasZgetriBatched (hipblasHandle_t handle, const int n, hipblasDoubleComplex *const A[], const int lda, int *ipiv, hipblasDoubleComplex *const C[], const int ldc, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasSgels (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, float *A, const int lda, float *B, const int ldb, int *info, int *deviceInfo)
 GELS solves an overdetermined (or underdetermined) linear system defined by an m-by-n matrix A, and a corresponding matrix B, using the QR factorization computed by GEQRF (or the LQ factorization computed by "GELQF"). More...
 
hipblasStatus_t hipblasDgels (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, double *A, const int lda, double *B, const int ldb, int *info, int *deviceInfo)
 GELS solves an overdetermined (or underdetermined) linear system defined by an m-by-n matrix A, and a corresponding matrix B, using the QR factorization computed by GEQRF (or the LQ factorization computed by "GELQF"). More...
 
hipblasStatus_t hipblasCgels (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, hipblasComplex *A, const int lda, hipblasComplex *B, const int ldb, int *info, int *deviceInfo)
 GELS solves an overdetermined (or underdetermined) linear system defined by an m-by-n matrix A, and a corresponding matrix B, using the QR factorization computed by GEQRF (or the LQ factorization computed by "GELQF"). More...
 
hipblasStatus_t hipblasZgels (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, hipblasDoubleComplex *A, const int lda, hipblasDoubleComplex *B, const int ldb, int *info, int *deviceInfo)
 GELS solves an overdetermined (or underdetermined) linear system defined by an m-by-n matrix A, and a corresponding matrix B, using the QR factorization computed by GEQRF (or the LQ factorization computed by "GELQF"). More...
 
hipblasStatus_t hipblasSgelsBatched (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, float *const A[], const int lda, float *const B[], const int ldb, int *info, int *deviceInfo, const int batchCount)
 gelsBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_BATCHED" (or the LQ factorizations computed by "GELQF_BATCHED"). More...
 
hipblasStatus_t hipblasDgelsBatched (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, double *const A[], const int lda, double *const B[], const int ldb, int *info, int *deviceInfo, const int batchCount)
 gelsBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_BATCHED" (or the LQ factorizations computed by "GELQF_BATCHED"). More...
 
hipblasStatus_t hipblasCgelsBatched (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, hipblasComplex *const A[], const int lda, hipblasComplex *const B[], const int ldb, int *info, int *deviceInfo, const int batchCount)
 gelsBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_BATCHED" (or the LQ factorizations computed by "GELQF_BATCHED"). More...
 
hipblasStatus_t hipblasZgelsBatched (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, hipblasDoubleComplex *const A[], const int lda, hipblasDoubleComplex *const B[], const int ldb, int *info, int *deviceInfo, const int batchCount)
 gelsBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_BATCHED" (or the LQ factorizations computed by "GELQF_BATCHED"). More...
 
hipblasStatus_t hipblasSgelsStridedBatched (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, float *A, const int lda, const hipblasStride strideA, float *B, const int ldb, const hipblasStride strideB, int *info, int *deviceInfo, const int batch_count)
 gelsStridedBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_STRIDED_BATCHED" (or the LQ factorizations computed by "GELQF_STRIDED_BATCHED"). More...
 
hipblasStatus_t hipblasDgelsStridedBatched (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, double *A, const int lda, const hipblasStride strideA, double *B, const int ldb, const hipblasStride strideB, int *info, int *deviceInfo, const int batch_count)
 gelsStridedBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_STRIDED_BATCHED" (or the LQ factorizations computed by "GELQF_STRIDED_BATCHED"). More...
 
hipblasStatus_t hipblasCgelsStridedBatched (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, hipblasComplex *A, const int lda, const hipblasStride strideA, hipblasComplex *B, const int ldb, const hipblasStride strideB, int *info, int *deviceInfo, const int batch_count)
 gelsStridedBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_STRIDED_BATCHED" (or the LQ factorizations computed by "GELQF_STRIDED_BATCHED"). More...
 
hipblasStatus_t hipblasZgelsStridedBatched (hipblasHandle_t handle, hipblasOperation_t trans, const int m, const int n, const int nrhs, hipblasDoubleComplex *A, const int lda, const hipblasStride strideA, hipblasDoubleComplex *B, const int ldb, const hipblasStride strideB, int *info, int *deviceInfo, const int batch_count)
 gelsStridedBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_STRIDED_BATCHED" (or the LQ factorizations computed by "GELQF_STRIDED_BATCHED"). More...
 
hipblasStatus_t hipblasSgeqrf (hipblasHandle_t handle, const int m, const int n, float *A, const int lda, float *ipiv, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasDgeqrf (hipblasHandle_t handle, const int m, const int n, double *A, const int lda, double *ipiv, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasCgeqrf (hipblasHandle_t handle, const int m, const int n, hipblasComplex *A, const int lda, hipblasComplex *ipiv, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasZgeqrf (hipblasHandle_t handle, const int m, const int n, hipblasDoubleComplex *A, const int lda, hipblasDoubleComplex *ipiv, int *info)
 SOLVER API. More...
 
hipblasStatus_t hipblasSgeqrfBatched (hipblasHandle_t handle, const int m, const int n, float *const A[], const int lda, float *const ipiv[], int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasDgeqrfBatched (hipblasHandle_t handle, const int m, const int n, double *const A[], const int lda, double *const ipiv[], int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasCgeqrfBatched (hipblasHandle_t handle, const int m, const int n, hipblasComplex *const A[], const int lda, hipblasComplex *const ipiv[], int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasZgeqrfBatched (hipblasHandle_t handle, const int m, const int n, hipblasDoubleComplex *const A[], const int lda, hipblasDoubleComplex *const ipiv[], int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasSgeqrfStridedBatched (hipblasHandle_t handle, const int m, const int n, float *A, const int lda, const hipblasStride strideA, float *ipiv, const hipblasStride strideP, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasDgeqrfStridedBatched (hipblasHandle_t handle, const int m, const int n, double *A, const int lda, const hipblasStride strideA, double *ipiv, const hipblasStride strideP, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasCgeqrfStridedBatched (hipblasHandle_t handle, const int m, const int n, hipblasComplex *A, const int lda, const hipblasStride strideA, hipblasComplex *ipiv, const hipblasStride strideP, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasZgeqrfStridedBatched (hipblasHandle_t handle, const int m, const int n, hipblasDoubleComplex *A, const int lda, const hipblasStride strideA, hipblasDoubleComplex *ipiv, const hipblasStride strideP, int *info, const int batchCount)
 SOLVER API. More...
 
hipblasStatus_t hipblasDotEx (hipblasHandle_t handle, int n, const void *x, hipblasDatatype_t xType, int incx, const void *y, hipblasDatatype_t yType, int incy, void *result, hipblasDatatype_t resultType, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasDotcEx (hipblasHandle_t handle, int n, const void *x, hipblasDatatype_t xType, int incx, const void *y, hipblasDatatype_t yType, int incy, void *result, hipblasDatatype_t resultType, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasDotBatchedEx (hipblasHandle_t handle, int n, const void *x, hipblasDatatype_t xType, int incx, const void *y, hipblasDatatype_t yType, int incy, int batchCount, void *result, hipblasDatatype_t resultType, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasDotcBatchedEx (hipblasHandle_t handle, int n, const void *x, hipblasDatatype_t xType, int incx, const void *y, hipblasDatatype_t yType, int incy, int batchCount, void *result, hipblasDatatype_t resultType, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasDotStridedBatchedEx (hipblasHandle_t handle, int n, const void *x, hipblasDatatype_t xType, int incx, hipblasStride stridex, const void *y, hipblasDatatype_t yType, int incy, hipblasStride stridey, int batchCount, void *result, hipblasDatatype_t resultType, hipblasDatatype_t executionType)
 BLAS EX API. More...
 
hipblasStatus_t hipblasDotcStridedBatchedEx (hipblasHandle_t handle, int n, const void *x, hipblasDatatype_t xType, int incx, hipblasStride stridex, const void *y, hipblasDatatype_t yType, int incy, hipblasStride stridey, int batchCount, void *result, hipblasDatatype_t resultType, hipblasDatatype_t executionType)
 BLAS EX API. More...
 

Macro Definition Documentation

◆ HIPBLAS_CLANG_STATIC

#define HIPBLAS_CLANG_STATIC

HIP = Heterogeneous-compute Interface for Portability

Define a extremely thin runtime layer that allows source code to be compiled unmodified through either AMD HCC or NVCC. Key features tend to be in the spirit and terminology of CUDA, but with a portable path to other accelerators as well.

This is the master include file for hipblas, wrapping around rocblas and cublas "version 2"

Typedef Documentation

◆ hipblasHalf

typedef uint16_t hipblasHalf

To specify the datatype to be unsigned short.

◆ hipblasHandle_t

typedef void* hipblasHandle_t

hipblasHanlde_t is a void pointer, to store the library context (either rocBLAS or cuBLAS)

◆ hipblasInt8

typedef int8_t hipblasInt8

To specify the datatype to be signed char.

◆ hipblasStride

typedef int64_t hipblasStride

Stride between matrices or vectors in strided_batched functions.

Enumeration Type Documentation

◆ hipblasAtomicsMode_t

Indicates if atomics operations are allowed. Not allowing atomic operations may generally improve determinism and repeatability of results at a cost of performance.

Enumerator
HIPBLAS_ATOMICS_NOT_ALLOWED 

Algorithms will refrain from atomics where applicable.

HIPBLAS_ATOMICS_ALLOWED 

Algorithms will take advantage of atomics where applicable.

◆ hipblasDatatype_t

Indicates the precision width of data stored in a blas type.

Enumerator
HIPBLAS_R_16F 

16 bit floating point, real

HIPBLAS_R_32F 

32 bit floating point, real

HIPBLAS_R_64F 

64 bit floating point, real

HIPBLAS_C_16F 

16 bit floating point, complex

HIPBLAS_C_32F 

32 bit floating point, complex

HIPBLAS_C_64F 

64 bit floating point, complex

HIPBLAS_R_8I 

8 bit signed integer, real

HIPBLAS_R_8U 

8 bit unsigned integer, real

HIPBLAS_R_32I 

32 bit signed integer, real

HIPBLAS_R_32U 

32 bit unsigned integer, real

HIPBLAS_C_8I 

8 bit signed integer, complex

HIPBLAS_C_8U 

8 bit unsigned integer, complex

HIPBLAS_C_32I 

32 bit signed integer, complex

HIPBLAS_C_32U 

32 bit unsigned integer, complex

HIPBLAS_R_16B 

16 bit bfloat, real

HIPBLAS_C_16B 

16 bit bfloat, complex

HIPBLAS_DATATYPE_INVALID 

Invalid datatype value, do not use

◆ hipblasDiagType_t

It is used by the triangular matrix routines to specify whether the matrix is unit triangular.

Enumerator
HIPBLAS_DIAG_NON_UNIT 

Non-unit triangular.

HIPBLAS_DIAG_UNIT 

Unit triangular.

◆ hipblasFillMode_t

Used by the Hermitian, symmetric and triangular matrix routines to specify whether the upper or lower triangle is being referenced.

Enumerator
HIPBLAS_FILL_MODE_UPPER 

Upper triangle

HIPBLAS_FILL_MODE_LOWER 

Lower triangle

HIPBLAS_FILL_MODE_FULL 

◆ hipblasGemmAlgo_t

Indicates if layer is active with bitmask.

Enumerator
HIPBLAS_GEMM_DEFAULT 

enumerator rocblas_gemm_algo_standard

◆ hipblasOperation_t

Used to specify whether the matrix is to be transposed or not.

Enumerator
HIPBLAS_OP_N 

Operate with the matrix.

HIPBLAS_OP_T 

Operate with the transpose of the matrix.

HIPBLAS_OP_C 

Operate with the conjugate transpose of the matrix.

◆ hipblasPointerMode_t

Indicates if scalar pointers are on host or device. This is used for scalars alpha and beta and for scalar function return values.

Enumerator
HIPBLAS_POINTER_MODE_HOST 

Scalar values affected by this variable will be located on the host.

HIPBLAS_POINTER_MODE_DEVICE 

Scalar values affected by this variable will be located on the device.

◆ hipblasSideMode_t

Indicates the side matrix A is located relative to matrix B during multiplication.

Enumerator
HIPBLAS_SIDE_LEFT 

Multiply general matrix by symmetric, Hermitian or triangular matrix on the left.

HIPBLAS_SIDE_RIGHT 

Multiply general matrix by symmetric, Hermitian or triangular matrix on the right.

HIPBLAS_SIDE_BOTH 

◆ hipblasStatus_t

hipblas status codes definition

Enumerator
HIPBLAS_STATUS_SUCCESS 

Function succeeds

HIPBLAS_STATUS_NOT_INITIALIZED 

HIPBLAS library not initialized

HIPBLAS_STATUS_ALLOC_FAILED 

resource allocation failed

HIPBLAS_STATUS_INVALID_VALUE 

unsupported numerical value was passed to function

HIPBLAS_STATUS_MAPPING_ERROR 

access to GPU memory space failed

HIPBLAS_STATUS_EXECUTION_FAILED 

GPU program failed to execute

HIPBLAS_STATUS_INTERNAL_ERROR 

an internal HIPBLAS operation failed

HIPBLAS_STATUS_NOT_SUPPORTED 

function not implemented

HIPBLAS_STATUS_ARCH_MISMATCH 

architecture mismatch

HIPBLAS_STATUS_HANDLE_IS_NULLPTR 

hipBLAS handle is null pointer

HIPBLAS_STATUS_INVALID_ENUM 

unsupported enum value was passed to function

HIPBLAS_STATUS_UNKNOWN 

back-end returned an unsupported status code

Function Documentation

◆ hipblasAxpyBatchedEx()

hipblasStatus_t hipblasAxpyBatchedEx ( hipblasHandle_t  handle,
int  n,
const void *  alpha,
hipblasDatatype_t  alphaType,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
void *  y,
hipblasDatatype_t  yType,
int  incy,
int  batchCount,
hipblasDatatype_t  executionType 
)

BLAS EX API.

axpyBatchedEx computes constant alpha multiplied by vector x, plus vector y over a set of batched vectors.

y := alpha * x + y

  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]alphadevice pointer or host pointer to specify the scalar alpha.
[in]alphaType[hipblasDatatype_t] specifies the datatype of alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]yType[hipblasDatatype_t] specifies the datatype of each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasAxpyEx()

hipblasStatus_t hipblasAxpyEx ( hipblasHandle_t  handle,
int  n,
const void *  alpha,
hipblasDatatype_t  alphaType,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
void *  y,
hipblasDatatype_t  yType,
int  incy,
hipblasDatatype_t  executionType 
)

BLAS EX API.

axpyEx computes constant alpha multiplied by vector x, plus vector y

y := alpha * x + y

- Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]alphadevice pointer or host pointer to specify the scalar alpha.
[in]alphaType[hipblasDatatype_t] specifies the datatype of alpha.
[in]xdevice pointer storing vector x.
[in]xType[hipblasDatatype_t] specifies the datatype of vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]ydevice pointer storing vector y.
[in]yType[hipblasDatatype_t] specifies the datatype of vector y.
[in]incy[int] specifies the increment for the elements of y.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasAxpyStridedBatchedEx()

hipblasStatus_t hipblasAxpyStridedBatchedEx ( hipblasHandle_t  handle,
int  n,
const void *  alpha,
hipblasDatatype_t  alphaType,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
hipblasStride  stridex,
void *  y,
hipblasDatatype_t  yType,
int  incy,
hipblasStride  stridey,
int  batchCount,
hipblasDatatype_t  executionType 
)

BLAS EX API.

axpyStridedBatchedEx computes constant alpha multiplied by vector x, plus vector y over a set of strided batched vectors.

y := alpha * x + y

  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]alphadevice pointer or host pointer to specify the scalar alpha.
[in]alphaType[hipblasDatatype_t] specifies the datatype of alpha.
[in]xdevice pointer to the first vector x_1.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) to the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size, for a typical case this means stridex >= n * incx.
[in,out]ydevice pointer to the first vector y_1.
[in]yType[hipblasDatatype_t] specifies the datatype of each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) to the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size, for a typical case this means stridey >= n * incy.
[in]batchCount[int] number of instances in the batch.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasBfdot()

hipblasStatus_t hipblasBfdot ( hipblasHandle_t  handle,
int  n,
const hipblasBfloat16 x,
int  incx,
const hipblasBfloat16 y,
int  incy,
hipblasBfloat16 result 
)

BLAS Level 1 API.

dot(u) performs the dot product of vectors x and y

result = x * y;

dotc performs the dot product of the conjugate of complex vector x and complex vector y

result = conjugate (x) * y;
  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the dot product. return is 0.0 if n <= 0.

◆ hipblasBfdotBatched()

hipblasStatus_t hipblasBfdotBatched ( hipblasHandle_t  handle,
int  n,
const hipblasBfloat16 *const  x[],
int  incx,
const hipblasBfloat16 *const  y[],
int  incy,
int  batchCount,
hipblasBfloat16 result 
)

BLAS Level 1 API.

dotBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasBfdotStridedBatched()

hipblasStatus_t hipblasBfdotStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasBfloat16 x,
int  incx,
hipblasStride  stridex,
const hipblasBfloat16 y,
int  incy,
hipblasStride  stridey,
int  batchCount,
hipblasBfloat16 result 
)

BLAS Level 1 API.

dotStridedBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcStridedBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasCaxpy()

hipblasStatus_t hipblasCaxpy ( hipblasHandle_t  handle,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
hipblasComplex y,
int  incy 
)

BLAS Level 1 API.

axpy computes constant alpha multiplied by vector x, plus vector y

y := alpha * x + y
  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]alphadevice pointer or host pointer to specify the scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[out]ydevice pointer storing vector y.
[in,out]incy[int] specifies the increment for the elements of y.

◆ hipblasCaxpyBatched()

hipblasStatus_t hipblasCaxpyBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  x[],
int  incx,
hipblasComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

axpyBatched compute y := alpha * x + y over a set of batched vectors.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]alphaspecifies the scalar alpha.
[in]xpointer storing vector x on the GPU.
[in]incx[int] specifies the increment for the elements of x.
[out]ypointer storing vector y on the GPU.
[in,out]incy[int] specifies the increment for the elements of y.
[in]batchCount[int] number of instances in the batch

◆ hipblasCaxpyStridedBatched()

hipblasStatus_t hipblasCaxpyStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
hipblasComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

axpyStridedBatched compute y := alpha * x + y over a set of strided batched vectors.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int]
[in]alphaspecifies the scalar alpha.
[in]xpointer storing vector x on the GPU.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] specifies the increment between vectors of x.
[out]ypointer storing vector y on the GPU.
[in,out]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] specifies the increment between vectors of y.
[in]batchCount[int] number of instances in the batch

◆ hipblasCcopy()

hipblasStatus_t hipblasCcopy ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
hipblasComplex y,
int  incy 
)

BLAS Level 1 API.

copy copies each element x[i] into y[i], for i = 1 , ... , n

y := x,
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x to be copied to y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasCcopyBatched()

hipblasStatus_t hipblasCcopyBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex *const  x[],
int  incx,
hipblasComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

copyBatched copies each element x_i[j] into y_i[j], for j = 1 , ... , n; i = 1 , ... , batchCount

y_i := x_i,

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i to be copied to y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasCcopyStridedBatched()

hipblasStatus_t hipblasCcopyStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
hipblasComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

copyStridedBatched copies each element x_i[j] into y_i[j], for j = 1 , ... , n; i = 1 , ... , batchCount

y_i := x_i,

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i to be copied to y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increments for the elements of vectors x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of vectors y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stride_y, however the user should take care to ensure that stride_y is of appropriate size, for a typical case this means stride_y >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasCdgmm()

hipblasStatus_t hipblasCdgmm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const hipblasComplex AP,
int  lda,
const hipblasComplex x,
int  incx,
hipblasComplex CP,
int  ldc 
)

BLAS Level 3 API.

dgmm performs one of the matrix-matrix operations

C = A * diag(x) if side == HIPBLAS_SIDE_RIGHT
C = diag(x) * A if side == HIPBLAS_SIDE_LEFT

where C and A are m by n dimensional matrices. diag( x ) is a diagonal matrix and x is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between values of x
[in,out]CPdevice pointer storing matrix C.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasCdgmmBatched()

hipblasStatus_t hipblasCdgmmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  x[],
int  incx,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

dgmmBatched performs one of the batched matrix-matrix operations

C_i = A_i * diag(x_i) for i = 0, 1, ... batchCount-1 if side == HIPBLAS_SIDE_RIGHT
C_i = diag(x_i) * A_i for i = 0, 1, ... batchCount-1 if side == HIPBLAS_SIDE_LEFT

where C_i and A_i are m by n dimensional matrices. diag(x_i) is a diagonal matrix and x_i is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, n )
[in]lda[int] specifies the leading dimension of A_i.
[in]xdevice array of device pointers storing each vector x_i on the GPU. Each x_i is of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT
[in]incx[int] specifies the increment between values of x_i
[in,out]CPdevice array of device pointers storing each matrix C_i on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasCdgmmStridedBatched()

hipblasStatus_t hipblasCdgmmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
hipblasComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

dgmmStridedBatched performs one of the batched matrix-matrix operations

C_i = A_i * diag(x_i)   if side == HIPBLAS_SIDE_RIGHT   for i = 0, 1, ... batchCount-1
C_i = diag(x_i) * A_i   if side == HIPBLAS_SIDE_LEFT    for i = 0, 1, ... batchCount-1

where C_i and A_i are m by n dimensional matrices. diag(x_i) is a diagonal matrix and x_i is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, n )
[in]lda[int] specifies the leading dimension of A.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xpointer to the first vector x_0 on the GPU. Each x_i is of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT
[in]incx[int] specifies the increment between values of x
[in]stridex[hipblasStride] stride from the start of one vector(x_i) and the next one (x_i+1)
[in,out]CPdevice pointer to the first matrix C_0 on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasCdotc()

hipblasStatus_t hipblasCdotc ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
const hipblasComplex y,
int  incy,
hipblasComplex result 
)

BLAS Level 1 API.

dot(u) performs the dot product of vectors x and y

result = x * y;

dotc performs the dot product of the conjugate of complex vector x and complex vector y

result = conjugate (x) * y;
  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the dot product. return is 0.0 if n <= 0.

◆ hipblasCdotcBatched()

hipblasStatus_t hipblasCdotcBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex *const  y[],
int  incy,
int  batchCount,
hipblasComplex result 
)

BLAS Level 1 API.

dotBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasCdotcStridedBatched()

hipblasStatus_t hipblasCdotcStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount,
hipblasComplex result 
)

BLAS Level 1 API.

dotStridedBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcStridedBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasCdotu()

hipblasStatus_t hipblasCdotu ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
const hipblasComplex y,
int  incy,
hipblasComplex result 
)

BLAS Level 1 API.

dot(u) performs the dot product of vectors x and y

result = x * y;

dotc performs the dot product of the conjugate of complex vector x and complex vector y

result = conjugate (x) * y;
  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the dot product. return is 0.0 if n <= 0.

◆ hipblasCdotuBatched()

hipblasStatus_t hipblasCdotuBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex *const  y[],
int  incy,
int  batchCount,
hipblasComplex result 
)

BLAS Level 1 API.

dotBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasCdotuStridedBatched()

hipblasStatus_t hipblasCdotuStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount,
hipblasComplex result 
)

BLAS Level 1 API.

dotStridedBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcStridedBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasCgbmv()

hipblasStatus_t hipblasCgbmv ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex x,
int  incx,
const hipblasComplex beta,
hipblasComplex y,
int  incy 
)

BLAS Level 2 API.

gbmv performs one of the matrix-vector operations

y := alpha*A*x    + beta*y,   or
y := alpha*A**T*x + beta*y,   or
y := alpha*A**H*x + beta*y,

where alpha and beta are scalars, x and y are vectors and A is an m by n banded matrix with kl sub-diagonals and ku super-diagonals.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]kl[int] number of sub-diagonals of A
[in]ku[int] number of super-diagonals of A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing banded matrix A. Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of A. Must be >= (kl + ku + 1)
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasCgbmvBatched()

hipblasStatus_t hipblasCgbmvBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex beta,
hipblasComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

gbmvBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n banded matrix with kl sub-diagonals and ku super-diagonals, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of each matrix A_i
[in]n[int] number of columns of each matrix A_i
[in]kl[int] number of sub-diagonals of each A_i
[in]ku[int] number of super-diagonals of each A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each banded matrix A_i. Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of each A_i. Must be >= (kl + ku + 1)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasCgbmvStridedBatched()

hipblasStatus_t hipblasCgbmvStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex beta,
hipblasComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

gbmvStridedBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n banded matrix with kl sub-diagonals and ku super-diagonals, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]kl[int] number of sub-diagonals of A
[in]ku[int] number of super-diagonals of A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer to first banded matrix (A_1). Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of A. Must be >= (kl + ku + 1)
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xdevice pointer to first vector (x_1).
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer to first vector (y_1).
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (x_i+1)
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasCgeam()

hipblasStatus_t hipblasCgeam ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex beta,
const hipblasComplex BP,
int  ldb,
hipblasComplex CP,
int  ldc 
)

BLAS Level 3 API.

geam performs one of the matrix-matrix operations

C = alpha*op( A ) + beta*op( B ),

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are matrices, with op( A ) an m by n matrix, op( B ) an m by n matrix, and C an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPdevice pointer storing matrix B.
[in]ldb[int] specifies the leading dimension of B.
[in,out]CPdevice pointer storing matrix C.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasCgeamBatched()

hipblasStatus_t hipblasCgeamBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex beta,
const hipblasComplex *const  BP[],
int  ldb,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

geamBatched performs one of the batched matrix-matrix operations

C_i = alpha*op( A_i ) + beta*op( B_i )  for i = 0, 1, ... batchCount - 1

where alpha and beta are scalars, and op(A_i), op(B_i) and C_i are m by n matrices and op( X ) is one of

op( X ) = X      or
op( X ) = X**T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, k ), where k is m when transA == HIPBLAS_OP_N and is n when transA == HIPBLAS_OP_T.
[in]lda[int] specifies the leading dimension of A.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPdevice array of device pointers storing each matrix B_i on the GPU. Each B_i is of dimension ( ldb, k ), where k is m when transB == HIPBLAS_OP_N and is n when transB == HIPBLAS_OP_T.
[in]ldb[int] specifies the leading dimension of B.
[in,out]CPdevice array of device pointers storing each matrix C_i on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]batchCount[int] number of instances i in the batch.

◆ hipblasCgeamStridedBatched()

hipblasStatus_t hipblasCgeamStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex beta,
const hipblasComplex BP,
int  ldb,
hipblasStride  strideB,
hipblasComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

geamStridedBatched performs one of the batched matrix-matrix operations

C_i = alpha*op( A_i ) + beta*op( B_i )  for i = 0, 1, ... batchCount - 1

where alpha and beta are scalars, and op(A_i), op(B_i) and C_i are m by n matrices and op( X ) is one of

op( X ) = X      or
op( X ) = X**T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, k ), where k is m when transA == HIPBLAS_OP_N and is n when transA == HIPBLAS_OP_T.
[in]lda[int] specifies the leading dimension of A.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPpointer to the first matrix B_0 on the GPU. Each B_i is of dimension ( ldb, k ), where k is m when transB == HIPBLAS_OP_N and is n when transB == HIPBLAS_OP_T.
[in]ldb[int] specifies the leading dimension of B.
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in,out]CPpointer to the first matrix C_0 on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasCgels()

hipblasStatus_t hipblasCgels ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
hipblasComplex A,
const int  lda,
hipblasComplex B,
const int  ldb,
int *  info,
int *  deviceInfo 
)

GELS solves an overdetermined (or underdetermined) linear system defined by an m-by-n matrix A, and a corresponding matrix B, using the QR factorization computed by GEQRF (or the LQ factorization computed by "GELQF").

Depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A X = B & \: \text{not transposed, or}\\ A' X = B & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X is found by minimizing

\[ || B - A X || \quad \text{(or} \: || B - A' X ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X is chosen such that \(|| X ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : currently unsupported
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of matrix A.
[in]nint. n >= 0.
The number of columns of matrix A.
[in]nrhsint. nrhs >= 0.
The number of columns of matrices B and X; i.e., the columns on the right hand side.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the matrix A. On exit, the QR (or LQ) factorization of A as returned by "GEQRF" (or "GELQF").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrix A.
[in,out]Bpointer to type. Array on the GPU of dimension ldb*nrhs.
On entry, the matrix B. On exit, when info = 0, B is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrix B.
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int on the GPU.
If info = 0, successful exit. If info = i > 0, the solution could not be computed because input matrix A is rank deficient; the i-th diagonal element of its triangular factor is zero.

◆ hipblasCgelsBatched()

hipblasStatus_t hipblasCgelsBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
hipblasComplex *const  A[],
const int  lda,
hipblasComplex *const  B[],
const int  ldb,
int *  info,
int *  deviceInfo,
const int  batchCount 
)

gelsBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_BATCHED" (or the LQ factorizations computed by "GELQF_BATCHED").

For each instance in the batch, depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A_j X_j = B_j & \: \text{not transposed, or}\\ A_j' X_j = B_j & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X_j is found by minimizing

\[ || B_j - A_j X_j || \quad \text{(or} \: || B_j - A_j' X_j ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X_j is chosen such that \(|| X_j ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z Note that cuBLAS backend supports only the non-transpose operation and only solves over-determined systems (m >= n).
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of all matrices A_j in the batch.
[in]nint. n >= 0.
The number of columns of all matrices A_j in the batch.
[in]nrhsint. nrhs >= 0.
The number of columns of all matrices B_j and X_j in the batch; i.e., the columns on the right hand side.
[in,out]Aarray of pointer to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the matrices A_j. On exit, the QR (or LQ) factorizations of A_j as returned by "GEQRF_BATCHED" (or "GELQF_BATCHED").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_j.
[in,out]Barray of pointer to type. Each pointer points to an array on the GPU of dimension ldb*nrhs.
On entry, the matrices B_j. On exit, when info[j] = 0, B_j is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrices B_j.
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int. Array of batchCount integers on the GPU.
If deviceInfo[j] = 0, successful exit for solution of A_j. If deviceInfo[j] = i > 0, the solution of A_j could not be computed because input matrix A_j is rank deficient; the i-th diagonal element of its triangular factor is zero.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasCgelsStridedBatched()

hipblasStatus_t hipblasCgelsStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
hipblasComplex A,
const int  lda,
const hipblasStride  strideA,
hipblasComplex B,
const int  ldb,
const hipblasStride  strideB,
int *  info,
int *  deviceInfo,
const int  batch_count 
)

gelsStridedBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_STRIDED_BATCHED" (or the LQ factorizations computed by "GELQF_STRIDED_BATCHED").

For each instance in the batch, depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A_j X_j = B_j & \: \text{not transposed, or}\\ A_j' X_j = B_j & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X_j is found by minimizing

\[ || B_j - A_j X_j || \quad \text{(or} \: || B_j - A_j' X_j ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X_j is chosen such that \(|| X_j ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : currently unsupported
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of all matrices A_j in the batch.
[in]nint. n >= 0.
The number of columns of all matrices A_j in the batch.
[in]nrhsint. nrhs >= 0.
The number of columns of all matrices B_j and X_j in the batch; i.e., the columns on the right hand side.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the matrices A_j. On exit, the QR (or LQ) factorizations of A_j as returned by "GEQRF_STRIDED_BATCHED" (or "GELQF_STRIDED_BATCHED").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_j.
[in]strideAhipblasStride.
Stride from the start of one matrix A_j to the next one A_(j+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n
[in,out]Bpointer to type. Array on the GPU (the size depends on the value of strideB).
On entry, the matrices B_j. On exit, when info[j] = 0, each B_j is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrices B_j.
[in]strideBhipblasStride.
Stride from the start of one matrix B_j to the next one B_(j+1). There is no restriction for the value of strideB. Normal use case is strideB >= ldb*nrhs
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int. Array of batchCount integers on the GPU.
If deviceInfo[j] = 0, successful exit for solution of A_j. If deviceInfo[j] = i > 0, the solution of A_j could not be computed because input matrix A_j is rank deficient; the i-th diagonal element of its triangular factor is zero.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasCgemm()

hipblasStatus_t hipblasCgemm ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex BP,
int  ldb,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc 
)

BLAS Level 3 API.

gemm performs one of the matrix-matrix operations

C = alpha*op( A )*op( B ) + beta*C,

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are matrices, with op( A ) an m by k matrix, op( B ) a k by n matrix and C an m by n matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t]
      .
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] number or rows of matrices op( A ) and C
[in]n[int] number of columns of matrices op( B ) and C
[in]k[int] number of columns of matrix op( A ) and number of rows of matrix op( B )
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]BPdevice pointer storing matrix B.
[in]ldb[int] specifies the leading dimension of B.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice pointer storing matrix C on the GPU.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasCgemmBatched()

hipblasStatus_t hipblasCgemmBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  BP[],
int  ldb,
const hipblasComplex beta,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

gemmBatched performs one of the batched matrix-matrix operations C_i = alpha*op( A_i )*op( B_i ) + beta*C_i, for i = 1, ..., batchCount. where op( X ) is one of op( X ) = X or op( X ) = X**T or op( X ) = X**H, alpha and beta are scalars, and A, B and C are strided batched matrices, with op( A ) an m by k by batchCount strided_batched matrix, op( B ) an k by n by batchCount strided_batched matrix and C an m by n by batchCount strided_batched matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimention m.
[in]n[int] matrix dimention n.
[in]k[int] matrix dimention k.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]BPdevice array of device pointers storing each matrix B_i.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice array of device pointers storing each matrix C_i.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]batchCount[int] number of gemm operations in the batch

◆ hipblasCgemmStridedBatched()

hipblasStatus_t hipblasCgemmStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
long long  strideA,
const hipblasComplex BP,
int  ldb,
long long  strideB,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc,
long long  strideC,
int  batchCount 
)

BLAS Level 3 API.

gemmStridedBatched performs one of the strided batched matrix-matrix operations

C_i = alpha*op( A_i )*op( B_i ) + beta*C_i, for i = 1, ..., batchCount.

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are strided batched matrices, with op( A ) an m by k by batchCount strided_batched matrix, op( B ) an k by n by batchCount strided_batched matrix and C an m by n by batchCount strided_batched matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimention m.
[in]n[int] matrix dimention n.
[in]k[int] matrix dimention k.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer pointing to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in]BPdevice pointer pointing to the first matrix B_1.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]strideB[hipblasStride] stride from the start of one B_i matrix to the next B_(i + 1).
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice pointer pointing to the first matrix C_1.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]strideC[hipblasStride] stride from the start of one C_i matrix to the next C_(i + 1).
[in]batchCount[int] number of gemm operatons in the batch

◆ hipblasCgemv()

hipblasStatus_t hipblasCgemv ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex x,
int  incx,
const hipblasComplex beta,
hipblasComplex y,
int  incy 
)

BLAS Level 2 API.

gemv performs one of the matrix-vector operations

y := alpha*A*x    + beta*y,   or
y := alpha*A**T*x + beta*y,   or
y := alpha*A**H*x + beta*y,

where alpha and beta are scalars, x and y are vectors and A is an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasCgemvBatched()

hipblasStatus_t hipblasCgemvBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex beta,
hipblasComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

gemvBatched performs a batch of matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrices A_i are tranposed (conjugated) or not
[in]m[int] number of rows of each matrix A_i
[in]n[int] number of columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each matrix A_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasCgemvStridedBatched()

hipblasStatus_t hipblasCgemvStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex beta,
hipblasComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

gemvStridedBatched performs a batch of matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] indicates whether matrices A_i are tranposed (conjugated) or not
[in]m[int] number of rows of matrices A_i
[in]n[int] number of columns of matrices A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer to the first matrix (A_1) in the batch.
[in]lda[int] specifies the leading dimension of matrices A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. When trans equals HIPBLAS_OP_N this typically means stridex >= n * incx, otherwise stridex >= m * incx.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of vectors y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. When trans equals HIPBLAS_OP_N this typically means stridey >= m * incy, otherwise stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasCgeqrf()

hipblasStatus_t hipblasCgeqrf ( hipblasHandle_t  handle,
const int  m,
const int  n,
hipblasComplex A,
const int  lda,
hipblasComplex ipiv,
int *  info 
)

SOLVER API.

geqrf computes a QR factorization of a general m-by-n matrix A.

The factorization has the form

\[ A = Q\left[\begin{array}{c} R\\ 0 \end{array}\right] \]

where R is upper triangular (upper trapezoidal if m < n), and Q is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q = H_1H_2\cdots H_k, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_i\) is given by

\[ H_i = I - \text{ipiv}[i] \cdot v_i v_i' \]

where the first i-1 elements of the Householder vector \(v_i\) are zero, and \(v_i[i] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of the matrix A.
[in]nint. n >= 0.
The number of columns of the matrix A.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the m-by-n matrix to be factored. On exit, the elements on and above the diagonal contain the factor R; the elements below the diagonal are the last m - i elements of Householder vector v_i.
[in]ldaint. lda >= m.
Specifies the leading dimension of A.
[out]ipivpointer to type. Array on the GPU of dimension min(m,n).
The Householder scalars.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.

◆ hipblasCgeqrfBatched()

hipblasStatus_t hipblasCgeqrfBatched ( hipblasHandle_t  handle,
const int  m,
const int  n,
hipblasComplex *const  A[],
const int  lda,
hipblasComplex *const  ipiv[],
int *  info,
const int  batchCount 
)

SOLVER API.

geqrfBatched computes the QR factorization of a batch of general m-by-n matrices.

The factorization of matrix \(A_i\) in the batch has the form

\[ A_i = Q_i\left[\begin{array}{c} R_i\\ 0 \end{array}\right] \]

where \(R_i\) is upper triangular (upper trapezoidal if m < n), and \(Q_i\) is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q_i = H_{i_1}H_{i_2}\cdots H_{i_k}, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_{i_j}\) is given by

\[ H_{i_j} = I - \text{ipiv}_i[j] \cdot v_{i_j} v_{i_j}' \]

where the first j-1 elements of Householder vector \(v_{i_j}\) are zero, and \(v_{i_j}[j] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of all the matrices A_i in the batch.
[in]nint. n >= 0.
The number of columns of all the matrices A_i in the batch.
[in,out]AArray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the m-by-n matrices A_i to be factored. On exit, the elements on and above the diagonal contain the factor R_i. The elements below the diagonal are the last m - j elements of Householder vector v_(i_j).
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_i.
[out]ipivarray of pointers to type. Each pointer points to an array on the GPU of dimension min(m, n).
Contains the vectors ipiv_i of corresponding Householder scalars.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasCgeqrfStridedBatched()

hipblasStatus_t hipblasCgeqrfStridedBatched ( hipblasHandle_t  handle,
const int  m,
const int  n,
hipblasComplex A,
const int  lda,
const hipblasStride  strideA,
hipblasComplex ipiv,
const hipblasStride  strideP,
int *  info,
const int  batchCount 
)

SOLVER API.

geqrfStridedBatched computes the QR factorization of a batch of general m-by-n matrices.

The factorization of matrix \(A_i\) in the batch has the form

\[ A_i = Q_i\left[\begin{array}{c} R_i\\ 0 \end{array}\right] \]

where \(R_i\) is upper triangular (upper trapezoidal if m < n), and \(Q_i\) is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q_i = H_{i_1}H_{i_2}\cdots H_{i_k}, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_{i_j}\) is given by

\[ H_{i_j} = I - \text{ipiv}_j[j] \cdot v_{i_j} v_{i_j}' \]

where the first j-1 elements of Householder vector \(v_{i_j}\) are zero, and \(v_{i_j}[j] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of all the matrices A_i in the batch.
[in]nint. n >= 0.
The number of columns of all the matrices A_i in the batch.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the m-by-n matrices A_i to be factored. On exit, the elements on and above the diagonal contain the factor R_i. The elements below the diagonal are the last m - j elements of Householder vector v_(i_j).
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n.
[out]ipivpointer to type. Array on the GPU (the size depends on the value of strideP).
Contains the vectors ipiv_i of corresponding Householder scalars.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use is strideP >= min(m,n).
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasCgerc()

hipblasStatus_t hipblasCgerc ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
const hipblasComplex y,
int  incy,
hipblasComplex AP,
int  lda 
)

BLAS Level 2 API.

ger,geru,gerc performs the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where alpha is a scalar, x and y are vectors, and A is an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of the matrix A.
[in]n[int] the number of columns of the matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasCgercBatched()

hipblasStatus_t hipblasCgercBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex *const  y[],
int  incy,
hipblasComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

gerBatched,geruBatched,gercBatched performs a batch of the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of eaceh matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasCgercStridedBatched()

hipblasStatus_t hipblasCgercStridedBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex y,
int  incy,
hipblasStride  stridey,
hipblasComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

gerStridedBatched,geruStridedBatched,gercStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**T, OR
A_i := A_i + alpha*x_i*y_i**H  for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increments for the elements of each vector x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size, for a typical case this means stridex >= m * incx.
[in,out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size, for a typical case this means stridey >= n * incy.
[in,out]APdevice pointer to the first matrix (A_1) in the batch.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasCgeru()

hipblasStatus_t hipblasCgeru ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
const hipblasComplex y,
int  incy,
hipblasComplex AP,
int  lda 
)

BLAS Level 2 API.

ger,geru,gerc performs the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where alpha is a scalar, x and y are vectors, and A is an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of the matrix A.
[in]n[int] the number of columns of the matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasCgeruBatched()

hipblasStatus_t hipblasCgeruBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex *const  y[],
int  incy,
hipblasComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

gerBatched,geruBatched,gercBatched performs a batch of the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of eaceh matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasCgeruStridedBatched()

hipblasStatus_t hipblasCgeruStridedBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex y,
int  incy,
hipblasStride  stridey,
hipblasComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

gerStridedBatched,geruStridedBatched,gercStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**T, OR
A_i := A_i + alpha*x_i*y_i**H  for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increments for the elements of each vector x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size, for a typical case this means stridex >= m * incx.
[in,out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size, for a typical case this means stridey >= n * incy.
[in,out]APdevice pointer to the first matrix (A_1) in the batch.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasCgetrf()

hipblasStatus_t hipblasCgetrf ( hipblasHandle_t  handle,
const int  n,
hipblasComplex A,
const int  lda,
int *  ipiv,
int *  info 
)

SOLVER API.

getrf computes the LU factorization of a general n-by-n matrix A using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization has the form:

\[ A = PLU \]

where P is a permutation matrix, L is lower triangular with unit diagonal elements, and U is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A = LU \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of the matrix A.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the n-by-n matrix A to be factored. On exit, the factors L and U from the factorization. The unit diagonal elements of L are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of A.
[out]ipivpointer to int. Array on the GPU of dimension n.
The vector of pivot indices. Elements of ipiv are 1-based indices. For 1 <= i <= n, the row i of the matrix was interchanged with row ipiv[i]. Matrix P of the factorization can be derived from ipiv. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[out]infopointer to a int on the GPU.
If info = 0, successful exit. If info = j > 0, U is singular. U[j,j] is the first zero pivot.

◆ hipblasCgetrfBatched()

hipblasStatus_t hipblasCgetrfBatched ( hipblasHandle_t  handle,
const int  n,
hipblasComplex *const  A[],
const int  lda,
int *  ipiv,
int *  info,
const int  batchCount 
)

SOLVER API.

getrfBatched computes the LU factorization of a batch of general n-by-n matrices using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization of matrix \(A_i\) in the batch has the form:

\[ A_i = P_iL_iU_i \]

where \(P_i\) is a permutation matrix, \(L_i\) is lower triangular with unit diagonal elements, and \(U_i\) is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A_i = L_iU_i \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of all matrices A_i in the batch.
[in,out]Aarray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the n-by-n matrices A_i to be factored. On exit, the factors L_i and U_i from the factorizations. The unit diagonal elements of L_i are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[out]ipivpointer to int. Array on the GPU.
Contains the vectors of pivot indices ipiv_i (corresponding to A_i). Dimension of ipiv_i is n. Elements of ipiv_i are 1-based indices. For each instance A_i in the batch and for 1 <= j <= n, the row j of the matrix A_i was interchanged with row ipiv_i[j]. Matrix P_i of the factorization can be derived from ipiv_i. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for factorization of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasCgetrfStridedBatched()

hipblasStatus_t hipblasCgetrfStridedBatched ( hipblasHandle_t  handle,
const int  n,
hipblasComplex A,
const int  lda,
const hipblasStride  strideA,
int *  ipiv,
const hipblasStride  strideP,
int *  info,
const int  batchCount 
)

SOLVER API.

getrfStridedBatched computes the LU factorization of a batch of general n-by-n matrices using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization of matrix \(A_i\) in the batch has the form:

\[ A_i = P_iL_iU_i \]

where \(P_i\) is a permutation matrix, \(L_i\) is lower triangular with unit diagonal elements, and \(U_i\) is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A_i = L_iU_i \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of all matrices A_i in the batch.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the n-by-n matrices A_i to be factored. On exit, the factors L_i and U_i from the factorization. The unit diagonal elements of L_i are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n
[out]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
Contains the vectors of pivots indices ipiv_i (corresponding to A_i). Dimension of ipiv_i is n. Elements of ipiv_i are 1-based indices. For each instance A_i in the batch and for 1 <= j <= n, the row j of the matrix A_i was interchanged with row ipiv_i[j]. Matrix P_i of the factorization can be derived from ipiv_i. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use case is strideP >= n.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for factorization of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasCgetriBatched()

hipblasStatus_t hipblasCgetriBatched ( hipblasHandle_t  handle,
const int  n,
hipblasComplex *const  A[],
const int  lda,
int *  ipiv,
hipblasComplex *const  C[],
const int  ldc,
int *  info,
const int  batchCount 
)

SOLVER API.

getriBatched computes the inverse \(C_i = A_i^{-1}\) of a batch of general n-by-n matrices \(A_i\).

The inverse is computed by solving the linear system

\[ A_i C_i = I \]

where I is the identity matrix, and \(A_i\) is factorized as \(A_i = P_i L_i U_i\) as given by getrfBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of rows and columns of all matrices A_i in the batch.
[in]Aarray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfBatched.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[in]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
The pivot indices returned by getrfBatched. ipiv can be passed in as a nullptr, this will assume that getrfBatched was called without partial pivoting.
[out]Carray of pointers to type. Each pointer points to an array on the GPU of dimension ldc*n.
If info[i] = 0, the inverse of matrices A_i. Otherwise, undefined.
[in]ldcint. ldc >= n.
Specifies the leading dimension of C_i.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for inversion of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasCgetrs()

hipblasStatus_t hipblasCgetrs ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
hipblasComplex A,
const int  lda,
const int *  ipiv,
hipblasComplex B,
const int  ldb,
int *  info 
)

SOLVER API.

getrs solves a system of n linear equations on n variables in its factorized form.

It solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A X = B & \: \text{not transposed,}\\ A^T X = B & \: \text{transposed, or}\\ A^H X = B & \: \text{conjugate transposed.} \end{array} \]

Matrix A is defined by its triangular factors as returned by getrf.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of A.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of the matrix B.
[in]Apointer to type. Array on the GPU of dimension lda*n.
The factors L and U of the factorization A = P*L*U returned by getrf.
[in]ldaint. lda >= n.
The leading dimension of A.
[in]ipivpointer to int. Array on the GPU of dimension n.
The pivot indices returned by getrf.
[in,out]Bpointer to type. Array on the GPU of dimension ldb*nrhs.
On entry, the right hand side matrix B. On exit, the solution matrix X.
[in]ldbint. ldb >= n.
The leading dimension of B.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.

◆ hipblasCgetrsBatched()

hipblasStatus_t hipblasCgetrsBatched ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
hipblasComplex *const  A[],
const int  lda,
const int *  ipiv,
hipblasComplex *const  B[],
const int  ldb,
int *  info,
const int  batchCount 
)

SOLVER API.

getrsBatched solves a batch of systems of n linear equations on n variables in its factorized forms.

For each instance i in the batch, it solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A_i X_i = B_i & \: \text{not transposed,}\\ A_i^T X_i = B_i & \: \text{transposed, or}\\ A_i^H X_i = B_i & \: \text{conjugate transposed.} \end{array} \]

Matrix \(A_i\) is defined by its triangular factors as returned by getrfBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations of each instance in the batch.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of all A_i matrices.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of all the matrices B_i.
[in]AArray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfBatched.
[in]ldaint. lda >= n.
The leading dimension of matrices A_i.
[in]ipivpointer to int. Array on the GPU.
Contains the vectors ipiv_i of pivot indices returned by getrfBatched.
[in,out]BArray of pointers to type. Each pointer points to an array on the GPU of dimension ldb*nrhs.
On entry, the right hand side matrices B_i. On exit, the solution matrix X_i of each system in the batch.
[in]ldbint. ldb >= n.
The leading dimension of matrices B_i.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of instances (systems) in the batch.

◆ hipblasCgetrsStridedBatched()

hipblasStatus_t hipblasCgetrsStridedBatched ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
hipblasComplex A,
const int  lda,
const hipblasStride  strideA,
const int *  ipiv,
const hipblasStride  strideP,
hipblasComplex B,
const int  ldb,
const hipblasStride  strideB,
int *  info,
const int  batchCount 
)

SOLVER API.

getrsStridedBatched solves a batch of systems of n linear equations on n variables in its factorized forms.

For each instance i in the batch, it solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A_i X_i = B_i & \: \text{not transposed,}\\ A_i^T X_i = B_i & \: \text{transposed, or}\\ A_i^H X_i = B_i & \: \text{conjugate transposed.} \end{array} \]

Matrix \(A_i\) is defined by its triangular factors as returned by getrfStridedBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations of each instance in the batch.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of all A_i matrices.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of all the matrices B_i.
[in]Apointer to type. Array on the GPU (the size depends on the value of strideA).
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfStridedBatched.
[in]ldaint. lda >= n.
The leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n.
[in]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
Contains the vectors ipiv_i of pivot indices returned by getrfStridedBatched.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use case is strideP >= n.
[in,out]Bpointer to type. Array on the GPU (size depends on the value of strideB).
On entry, the right hand side matrices B_i. On exit, the solution matrix X_i of each system in the batch.
[in]ldbint. ldb >= n.
The leading dimension of matrices B_i.
[in]strideBhipblasStride.
Stride from the start of one matrix B_i to the next one B_(i+1). There is no restriction for the value of strideB. Normal use case is strideB >= ldb*nrhs.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of instances (systems) in the batch.

◆ hipblasChbmv()

hipblasStatus_t hipblasChbmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex x,
int  incx,
const hipblasComplex beta,
hipblasComplex y,
int  incy 
)

BLAS Level 2 API.

hbmv performs the matrix-vector operations

y := alpha*A*x + beta*y

where alpha and beta are scalars, x and y are n element vectors and A is an n by n Hermitian band matrix, with k super-diagonals.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is being supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is being supplied.
[in]n[int] the order of the matrix A.
[in]k[int] the number of super-diagonals of the matrix A. Must be >= 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing matrix A. Of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The leading (k + 1) by n part of A must contain the upper triangular band part of the Hermitian matrix, with the leading diagonal in row (k + 1), the first super-diagonal on the RHS of row k, etc. The top left k by x triangle of A will not be referenced. Ex (upper, lda = n = 4, k = 1): A Represented matrix (0,0) (5,9) (6,8) (7,7) (1, 0) (5, 9) (0, 0) (0, 0) (1,0) (2,0) (3,0) (4,0) (5,-9) (2, 0) (6, 8) (0, 0) (0,0) (0,0) (0,0) (0,0) (0, 0) (6,-8) (3, 0) (7, 7) (0,0) (0,0) (0,0) (0,0) (0, 0) (0, 0) (7,-7) (4, 0)

if uplo == HIPBLAS_FILL_MODE_LOWER: The leading (k + 1) by n part of A must contain the lower triangular band part of the Hermitian matrix, with the leading diagonal in row (1), the first sub-diagonal on the LHS of row 2, etc. The bottom right k by k triangle of A will not be referenced. Ex (lower, lda = 2, n = 4, k = 1): A Represented matrix (1,0) (2,0) (3,0) (4,0) (1, 0) (5,-9) (0, 0) (0, 0) (5,9) (6,8) (7,7) (0,0) (5, 9) (2, 0) (6,-8) (0, 0) (0, 0) (6, 8) (3, 0) (7,-7) (0, 0) (0, 0) (7, 7) (4, 0)

As a Hermitian matrix, the imaginary part of the main diagonal of A will not be referenced and is assumed to be == 0.

Parameters
[in]lda[int] specifies the leading dimension of A. must be >= k + 1
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasChbmvBatched()

hipblasStatus_t hipblasChbmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex beta,
hipblasComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

hbmvBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian band matrix with k super-diagonals, for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is being supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is being supplied.
[in]n[int] the order of each matrix A_i.
[in]k[int] the number of super-diagonals of each matrix A_i. Must be >= 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The leading (k + 1) by n part of each A_i must contain the upper triangular band part of the Hermitian matrix, with the leading diagonal in row (k + 1), the first super-diagonal on the RHS of row k, etc. The top left k by x triangle of each A_i will not be referenced. Ex (upper, lda = n = 4, k = 1): A Represented matrix (0,0) (5,9) (6,8) (7,7) (1, 0) (5, 9) (0, 0) (0, 0) (1,0) (2,0) (3,0) (4,0) (5,-9) (2, 0) (6, 8) (0, 0) (0,0) (0,0) (0,0) (0,0) (0, 0) (6,-8) (3, 0) (7, 7) (0,0) (0,0) (0,0) (0,0) (0, 0) (0, 0) (7,-7) (4, 0)

if uplo == HIPBLAS_FILL_MODE_LOWER: The leading (k + 1) by n part of each A_i must contain the lower triangular band part of the Hermitian matrix, with the leading diagonal in row (1), the first sub-diagonal on the LHS of row 2, etc. The bottom right k by k triangle of each A_i will not be referenced. Ex (lower, lda = 2, n = 4, k = 1): A Represented matrix (1,0) (2,0) (3,0) (4,0) (1, 0) (5,-9) (0, 0) (0, 0) (5,9) (6,8) (7,7) (0,0) (5, 9) (2, 0) (6,-8) (0, 0) (0, 0) (6, 8) (3, 0) (7,-7) (0, 0) (0, 0) (7, 7) (4, 0)

As a Hermitian matrix, the imaginary part of the main diagonal of each A_i will not be referenced and is assumed to be == 0.

Parameters
[in]lda[int] specifies the leading dimension of each A_i. must be >= max(1, n)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of y.
[in]batchCount[int] number of instances in the batch.

◆ hipblasChbmvStridedBatched()

hipblasStatus_t hipblasChbmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex beta,
hipblasComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

hbmvStridedBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian band matrix with k super-diagonals, for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is being supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is being supplied.
[in]n[int] the order of each matrix A_i.
[in]k[int] the number of super-diagonals of each matrix A_i. Must be >= 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array pointing to the first matrix A_1. Each A_i is of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The leading (k + 1) by n part of each A_i must contain the upper triangular band part of the Hermitian matrix, with the leading diagonal in row (k + 1), the first super-diagonal on the RHS of row k, etc. The top left k by x triangle of each A_i will not be referenced. Ex (upper, lda = n = 4, k = 1): A Represented matrix (0,0) (5,9) (6,8) (7,7) (1, 0) (5, 9) (0, 0) (0, 0) (1,0) (2,0) (3,0) (4,0) (5,-9) (2, 0) (6, 8) (0, 0) (0,0) (0,0) (0,0) (0,0) (0, 0) (6,-8) (3, 0) (7, 7) (0,0) (0,0) (0,0) (0,0) (0, 0) (0, 0) (7,-7) (4, 0)

if uplo == HIPBLAS_FILL_MODE_LOWER: The leading (k + 1) by n part of each A_i must contain the lower triangular band part of the Hermitian matrix, with the leading diagonal in row (1), the first sub-diagonal on the LHS of row 2, etc. The bottom right k by k triangle of each A_i will not be referenced. Ex (lower, lda = 2, n = 4, k = 1): A Represented matrix (1,0) (2,0) (3,0) (4,0) (1, 0) (5,-9) (0, 0) (0, 0) (5,9) (6,8) (7,7) (0,0) (5, 9) (2, 0) (6,-8) (0, 0) (0, 0) (6, 8) (3, 0) (7,-7) (0, 0) (0, 0) (7, 7) (4, 0)

As a Hermitian matrix, the imaginary part of the main diagonal of each A_i will not be referenced and is assumed to be == 0.

Parameters
[in]lda[int] specifies the leading dimension of each A_i. must be >= max(1, n)
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xdevice array pointing to the first vector y_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array pointing to the first vector y_1.
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasChemm()

hipblasStatus_t hipblasChemm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex BP,
int  ldb,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc 
)

BLAS Level 3 API.

hemm performs one of the matrix-matrix operations:

C := alpha*A*B + beta*C if side == HIPBLAS_SIDE_LEFT, C := alpha*B*A + beta*C if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B and C are m by n matrices, and A is a Hermitian matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C := alpha*A*B + beta*C HIPBLAS_SIDE_RIGHT: C := alpha*B*A + beta*C
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix
[in]n[int] n specifies the number of rows of B and C. n >= 0.
[in]k[int] n specifies the number of columns of B and C. k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A and B are not referenced.
[in]APpointer storing matrix A on the GPU. A is m by m if side == HIPBLAS_SIDE_LEFT A is n by n if side == HIPBLAS_SIDE_RIGHT Only the upper/lower triangular part is accessed. The imaginary component of the diagonal elements is not used.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPpointer storing matrix B on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m )

◆ hipblasChemmBatched()

hipblasStatus_t hipblasChemmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  BP[],
int  ldb,
const hipblasComplex beta,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

hemmBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a Hermitian matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]n[int] n specifies the number of rows of B_i and C_i. n >= 0.
[in]k[int] k specifies the number of columns of B_i and C_i. k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT Only the upper/lower triangular part is accessed. The imaginary component of the diagonal elements is not used.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPdevice array of device pointers storing each matrix B_i on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C_i need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C_i. ldc >= max( 1, m )
[in]batchCount[int] number of instances in the batch.

◆ hipblasChemmStridedBatched()

hipblasStatus_t hipblasChemmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex BP,
int  ldb,
hipblasStride  strideB,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

hemmStridedBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a Hermitian matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]n[int] n specifies the number of rows of B_i and C_i. n >= 0.
[in]k[int] k specifies the number of columns of B_i and C_i. k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice pointer to first matrix A_1 A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT Only the upper/lower triangular part is accessed. The imaginary component of the diagonal elements is not used.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPdevice pointer to first matrix B_1 of dimension (ldb, n) on the GPU
[in]ldb[int] ldb specifies the first dimension of B_i. if side = HIPBLAS_OP_N, ldb >= max( 1, m ), otherwise ldb >= max( 1, n ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice pointer to first matrix C_1 of dimension (ldc, n) on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m )
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasChemv()

hipblasStatus_t hipblasChemv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex x,
int  incx,
const hipblasComplex beta,
hipblasComplex y,
int  incy 
)

BLAS Level 2 API.

hemv performs one of the matrix-vector operations

y := alpha*A*x + beta*y

where alpha and beta are scalars, x and y are n element vectors and A is an n by n Hermitian matrix.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of the Hermitian matrix A is supplied. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of the Hermitian matrix A is supplied.
[in]n[int] the order of the matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing matrix A. Of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A must contain the upper triangular part of a Hermitian matrix. The lower triangular part of A will not be referenced. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A must contain the lower triangular part of a Hermitian matrix. The upper triangular part of A will not be referenced. As a Hermitian matrix, the imaginary part of the main diagonal of A will not be referenced and is assumed to be == 0.
[in]lda[int] specifies the leading dimension of A. must be >= max(1, n)
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasChemvBatched()

hipblasStatus_t hipblasChemvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex beta,
hipblasComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

hemvBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian matrix, for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of the Hermitian matrix A is supplied. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of the Hermitian matrix A is supplied.
[in]n[int] the order of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i must contain the upper triangular part of a Hermitian matrix. The lower triangular part of each A_i will not be referenced. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i must contain the lower triangular part of a Hermitian matrix. The upper triangular part of each A_i will not be referenced. As a Hermitian matrix, the imaginary part of the main diagonal of each A_i will not be referenced and is assumed to be == 0.
[in]lda[int] specifies the leading dimension of each A_i. must be >= max(1, n)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of y.
[in]batchCount[int] number of instances in the batch.

◆ hipblasChemvStridedBatched()

hipblasStatus_t hipblasChemvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex beta,
hipblasComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

hemvStridedBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian matrix, for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of the Hermitian matrix A is supplied. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of the Hermitian matrix A is supplied.
[in]n[int] the order of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i must contain the upper triangular part of a Hermitian matrix. The lower triangular part of each A_i will not be referenced. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i must contain the lower triangular part of a Hermitian matrix. The upper triangular part of each A_i will not be referenced. As a Hermitian matrix, the imaginary part of the main diagonal of each A_i will not be referenced and is assumed to be == 0.
[in]lda[int] specifies the leading dimension of each A_i. must be >= max(1, n)
[in]strideA[hipblasStride] stride from the start of one (A_i) to the next (A_i+1)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCher()

hipblasStatus_t hipblasCher ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const hipblasComplex x,
int  incx,
hipblasComplex AP,
int  lda 
)

BLAS Level 2 API.

her performs the matrix-vector operations

A := A + alpha*x*x**H

where alpha is a real scalar, x is a vector, and A is an n by n Hermitian matrix.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in A. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in A.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing the specified triangular portion of the Hermitian matrix A. Of size (lda * n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the Hermitian matrix A is supplied. The lower triangluar portion will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the Hermitian matrix A is supplied. The upper triangular portion will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of A. Must be at least max(1, n).

◆ hipblasCher2()

hipblasStatus_t hipblasCher2 ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
const hipblasComplex y,
int  incy,
hipblasComplex AP,
int  lda 
)

BLAS Level 2 API.

her2 performs the matrix-vector operations

A := A + alpha*x*y**H + conj(alpha)*y*x**H

where alpha is a complex scalar, x and y are vectors, and A is an n by n Hermitian matrix.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing the specified triangular portion of the Hermitian matrix A. Of size (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the Hermitian matrix A is supplied. The lower triangular portion of A will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the Hermitian matrix A is supplied. The upper triangular portion of A will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of A. Must be at least max(lda, 1).

◆ hipblasCher2Batched()

hipblasStatus_t hipblasCher2Batched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex *const  y[],
int  incy,
hipblasComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

her2Batched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**H + conj(alpha)*y_i*x_i**H

where alpha is a complex scalar, x_i and y_i are vectors, and A_i is an n by n Hermitian matrix for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in,out]APdevice array of device pointers storing the specified triangular portion of each Hermitian matrix A_i of size (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The lower triangular portion of each A_i will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The upper triangular portion of each A_i will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of each A_i. Must be at least max(lda, 1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCher2k()

hipblasStatus_t hipblasCher2k ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex BP,
int  ldb,
const float *  beta,
hipblasComplex CP,
int  ldc 
)

BLAS Level 3 API.

her2k performs one of the matrix-matrix operations for a Hermitian rank-2k update

C := alpha*op( A )*op( B )^H + conj(alpha)*op( B )*op( A )^H + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrices, and C is a n x n Hermitian matrix stored as either upper or lower.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^H, op( B ) = B^H,  and A and B are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op( A ) = A^H, op( B ) = B^H HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasCher2kBatched()

hipblasStatus_t hipblasCher2kBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  BP[],
int  ldb,
const float *  beta,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

her2kBatched performs a batch of the matrix-matrix operations for a Hermitian rank-2k update

C_i := alpha*op( A_i )*op( B_i )^H + conj(alpha)*op( B_i )*op( A_i )^H + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrices, and C_i is a n x n Hermitian matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^H, op( B_i ) = B_i^H,  and A_i and B_i are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op(A) = A^H HIPBLAS_OP_N: op(A) = A
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCher2kStridedBatched()

hipblasStatus_t hipblasCher2kStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex BP,
int  ldb,
hipblasStride  strideB,
const float *  beta,
hipblasComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

her2kStridedBatched performs a batch of the matrix-matrix operations for a Hermitian rank-2k update

C_i := alpha*op( A_i )*op( B_i )^H + conj(alpha)*op( B_i )*op( A_i )^H + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrices, and C_i is a n x n Hermitian matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^H, op( B_i ) = B_i^H,  and A_i and B_i are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op( A_i ) = A_i^H, op( B_i ) = B_i^H HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasCher2StridedBatched()

hipblasStatus_t hipblasCher2StridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex y,
int  incy,
hipblasStride  stridey,
hipblasComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

her2StridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**H + conj(alpha)*y_i*x_i**H

where alpha is a complex scalar, x_i and y_i are vectors, and A_i is an n by n Hermitian matrix for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the stride between the beginning of one vector (x_i) and the next (x_i+1).
[in]ydevice pointer pointing to the first vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] specifies the stride between the beginning of one vector (y_i) and the next (y_i+1).
[in,out]APdevice pointer pointing to the first matrix (A_1). Stores the specified triangular portion of each Hermitian matrix A_i. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The lower triangular portion of each A_i will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The upper triangular portion of each A_i will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of each A_i. Must be at least max(lda, 1).
[in]strideA[hipblasStride] specifies the stride between the beginning of one matrix (A_i) and the next (A_i+1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCherBatched()

hipblasStatus_t hipblasCherBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const hipblasComplex *const  x[],
int  incx,
hipblasComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

herBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**H

where alpha is a real scalar, x_i is a vector, and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in A. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in A.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing the specified triangular portion of each Hermitian matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The lower triangular portion of each A_i will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The upper triangular portion of each A_i will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of each A_i. Must be at least max(1, n).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCherk()

hipblasStatus_t hipblasCherk ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const hipblasComplex AP,
int  lda,
const float *  beta,
hipblasComplex CP,
int  ldc 
)

BLAS Level 3 API.

herk performs one of the matrix-matrix operations for a Hermitian rank-k update

C := alpha*op( A )*op( A )^H + beta*C

where alpha and beta are scalars, op(A) is an n by k matrix, and C is a n x n Hermitian matrix stored as either upper or lower.

op( A ) = A,  and A is n by k if transA == HIPBLAS_OP_N
op( A ) = A^H and A is k by n if transA == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op(A) = A^H HIPBLAS_ON_N: op(A) = A
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if transA = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasCherkBatched()

hipblasStatus_t hipblasCherkBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const hipblasComplex *const  AP[],
int  lda,
const float *  beta,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

herkBatched performs a batch of the matrix-matrix operations for a Hermitian rank-k update

C_i := alpha*op( A_i )*op( A_i )^H + beta*C_i

where alpha and beta are scalars, op(A) is an n by k matrix, and C_i is a n x n Hermitian matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^H and A_i is k by n if transA == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op(A) = A^H HIPBLAS_OP_N: op(A) = A
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCherkStridedBatched()

hipblasStatus_t hipblasCherkStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const float *  beta,
hipblasComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

herkStridedBatched performs a batch of the matrix-matrix operations for a Hermitian rank-k update

C_i := alpha*op( A_i )*op( A_i )^H + beta*C_i

where alpha and beta are scalars, op(A) is an n by k matrix, and C_i is a n x n Hermitian matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^H and A_i is k by n if transA == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op(A) = A^H HIPBLAS_OP_N: op(A) = A
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasCherkx()

hipblasStatus_t hipblasCherkx ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex BP,
int  ldb,
const float *  beta,
hipblasComplex CP,
int  ldc 
)

BLAS Level 3 API.

herkx performs one of the matrix-matrix operations for a Hermitian rank-k update

C := alpha*op( A )*op( B )^H + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrices, and C is a n x n Hermitian matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A )*op( B )^T will be Hermitian.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^H, op( B ) = B^H,  and A and B are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op( A ) = A^H, op( B ) = B^H HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasCherkxBatched()

hipblasStatus_t hipblasCherkxBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  BP[],
int  ldb,
const float *  beta,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

herkxBatched performs a batch of the matrix-matrix operations for a Hermitian rank-k update

C_i := alpha*op( A_i )*op( B_i )^H + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrices, and C_i is a n x n Hermitian matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A )*op( B )^T will be Hermitian.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^H, op( B_i ) = B_i^H,  and A_i and B_i are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op(A) = A^H HIPBLAS_OP_N: op(A) = A
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCherkxStridedBatched()

hipblasStatus_t hipblasCherkxStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex BP,
int  ldb,
hipblasStride  strideB,
const float *  beta,
hipblasComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

herkxStridedBatched performs a batch of the matrix-matrix operations for a Hermitian rank-k update

C_i := alpha*op( A_i )*op( B_i )^H + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrices, and C_i is a n x n Hermitian matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A )*op( B )^T will be Hermitian.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^H, op( B_i ) = B_i^H,  and A_i and B_i are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op( A_i ) = A_i^H, op( B_i ) = B_i^H HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasCherStridedBatched()

hipblasStatus_t hipblasCherStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
hipblasComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

herStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**H

where alpha is a real scalar, x_i is a vector, and A_i is an n by n Hermitian matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in A. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in A.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in,out]APdevice array of device pointers storing the specified triangular portion of each Hermitian matrix A_i. Points to the first matrix (A_1). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The lower triangular portion of each A_i will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The upper triangular portion of each A_i will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasChpmv()

hipblasStatus_t hipblasChpmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
const hipblasComplex x,
int  incx,
const hipblasComplex beta,
hipblasComplex y,
int  incy 
)

BLAS Level 2 API.

hpmv performs the matrix-vector operation

y := alpha*A*x + beta*y

where alpha and beta are scalars, x and y are n element vectors and A is an n by n Hermitian matrix, supplied in packed form (see description below).

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of the Hermitian matrix A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of the Hermitian matrix A is supplied in AP.
[in]n[int] the order of the matrix A, must be >= 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing the packed version of the specified triangular portion of the Hermitian matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,1), (4,0), (3,2), (5,-1), (6,0)] (3,-2) (5, 1) (6, 0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,-1), (3,-2), (4,0), (5,1), (6,0)] (3,-2) (5, 1) (6, 0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasChpmvBatched()

hipblasStatus_t hipblasChpmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex beta,
hipblasComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

hpmvBatched performs the matrix-vector operation

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian matrix, supplied in packed form (see description below), for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of each Hermitian matrix A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of each Hermitian matrix A_i is supplied in AP.
[in]n[int] the order of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer of device pointers storing the packed version of the specified triangular portion of each Hermitian matrix A_i. Each A_i is of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that each AP_i contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,1), (4,0), (3,2), (5,-1), (6,0)] (3,-2) (5, 1) (6, 0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that each AP_i contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,-1), (3,-2), (4,0), (5,1), (6,0)] (3,-2) (5, 1) (6, 0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of y.
[in]batchCount[int] number of instances in the batch.

◆ hipblasChpmvStridedBatched()

hipblasStatus_t hipblasChpmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
hipblasStride  strideA,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex beta,
hipblasComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

hpmvStridedBatched performs the matrix-vector operation

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian matrix, supplied in packed form (see description below), for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of each Hermitian matrix A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of each Hermitian matrix A_i is supplied in AP.
[in]n[int] the order of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer pointing to the beginning of the first matrix (AP_1). Stores the packed version of the specified triangular portion of each Hermitian matrix AP_i of size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that each AP_i contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,1), (4,0), (3,2), (5,-1), (6,0)] (3,-2) (5, 1) (6, 0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that each AP_i contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,-1), (3,-2), (4,0), (5,1), (6,0)] (3,-2) (5, 1) (6, 0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]strideA[hipblasStride] stride from the start of one matrix (AP_i) and the next one (AP_i+1).
[in]xdevice array pointing to the beginning of the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array pointing to the beginning of the first vector (y_1).
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasChpr()

hipblasStatus_t hipblasChpr ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const hipblasComplex x,
int  incx,
hipblasComplex AP 
)

BLAS Level 2 API.

hpr performs the matrix-vector operations

A := A + alpha*x*x**H

where alpha is a real scalar, x is a vector, and A is an n by n Hermitian matrix, supplied in packed form.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in AP.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing the packed version of the specified triangular portion of the Hermitian matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.

◆ hipblasChpr2()

hipblasStatus_t hipblasChpr2 ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
const hipblasComplex y,
int  incy,
hipblasComplex AP 
)

BLAS Level 2 API.

hpr2 performs the matrix-vector operations

A := A + alpha*x*y**H + conj(alpha)*y*x**H

where alpha is a complex scalar, x and y are vectors, and A is an n by n Hermitian matrix, supplied in packed form.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in AP.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing the packed version of the specified triangular portion of the Hermitian matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.

◆ hipblasChpr2Batched()

hipblasStatus_t hipblasChpr2Batched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex *const  y[],
int  incy,
hipblasComplex *const  AP[],
int  batchCount 
)

BLAS Level 2 API.

hpr2Batched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**H + conj(alpha)*y_i*x_i**H

where alpha is a complex scalar, x_i and y_i are vectors, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each Hermitian matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]batchCount[int] number of instances in the batch.

◆ hipblasChpr2StridedBatched()

hipblasStatus_t hipblasChpr2StridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex y,
int  incy,
hipblasStride  stridey,
hipblasComplex AP,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

hpr2StridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**H + conj(alpha)*y_i*x_i**H

where alpha is a complex scalar, x_i and y_i are vectors, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in]ydevice pointer pointing to the first vector (y_1).
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1).
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each Hermitian matrix A_i. Points to the first matrix (A_1). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasChprBatched()

hipblasStatus_t hipblasChprBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const hipblasComplex *const  x[],
int  incx,
hipblasComplex *const  AP[],
int  batchCount 
)

BLAS Level 2 API.

hprBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**H

where alpha is a real scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each Hermitian matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]batchCount[int] number of instances in the batch.

◆ hipblasChprStridedBatched()

hipblasStatus_t hipblasChprStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
hipblasComplex AP,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

hprStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**H

where alpha is a real scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each Hermitian matrix A_i. Points to the first matrix (A_1). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasCreate()

hipblasStatus_t hipblasCreate ( hipblasHandle_t handle)

Create hipblas handle.

◆ hipblasCrot()

hipblasStatus_t hipblasCrot ( hipblasHandle_t  handle,
int  n,
hipblasComplex x,
int  incx,
hipblasComplex y,
int  incy,
const float *  c,
const hipblasComplex s 
)

BLAS Level 1 API.

rot applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to vectors x and y. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment between elements of y.
[in]cdevice pointer or host pointer storing scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer storing scalar sine component of the rotation matrix.

◆ hipblasCrotBatched()

hipblasStatus_t hipblasCrotBatched ( hipblasHandle_t  handle,
int  n,
hipblasComplex *const  x[],
int  incx,
hipblasComplex *const  y[],
int  incy,
const float *  c,
const hipblasComplex s,
int  batchCount 
)

BLAS Level 1 API.

rotBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice array of deivce pointers storing each vector x_i.
[in]incx[int] specifies the increment between elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment between elements of each y_i.
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasCrotg()

hipblasStatus_t hipblasCrotg ( hipblasHandle_t  handle,
hipblasComplex a,
hipblasComplex b,
float *  c,
hipblasComplex s 
)

BLAS Level 1 API.

rotg creates the Givens rotation matrix for the vector (a b). Scalars c and s and arrays a and b may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice pointer or host pointer to input vector element, overwritten with r.
[in,out]bdevice pointer or host pointer to input vector element, overwritten with z.
[in,out]cdevice pointer or host pointer to cosine element of Givens rotation.
[in,out]sdevice pointer or host pointer sine element of Givens rotation.

◆ hipblasCrotgBatched()

hipblasStatus_t hipblasCrotgBatched ( hipblasHandle_t  handle,
hipblasComplex *const  a[],
hipblasComplex *const  b[],
float *const  c[],
hipblasComplex *const  s[],
int  batchCount 
)

BLAS Level 1 API.

rotgBatched creates the Givens rotation matrix for the batched vectors (a_i b_i), for i = 1, ..., batchCount. a, b, c, and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice array of device pointers storing each single input vector element a_i, overwritten with r_i.
[in,out]bdevice array of device pointers storing each single input vector element b_i, overwritten with z_i.
[in,out]cdevice array of device pointers storing each cosine element of Givens rotation for the batch.
[in,out]sdevice array of device pointers storing each sine element of Givens rotation for the batch.
[in]batchCount[int] number of batches (length of arrays a, b, c, and s).

◆ hipblasCrotgStridedBatched()

hipblasStatus_t hipblasCrotgStridedBatched ( hipblasHandle_t  handle,
hipblasComplex a,
hipblasStride  stridea,
hipblasComplex b,
hipblasStride  strideb,
float *  c,
hipblasStride  stridec,
hipblasComplex s,
hipblasStride  strides,
int  batchCount 
)

BLAS Level 1 API.

rotgStridedBatched creates the Givens rotation matrix for the strided batched vectors (a_i b_i), for i = 1, ..., batchCount. a, b, c, and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice strided_batched pointer or host strided_batched pointer to first single input vector element a_1, overwritten with r.
[in]stridea[hipblasStride] distance between elements of a in batch (distance between a_i and a_(i + 1))
[in,out]bdevice strided_batched pointer or host strided_batched pointer to first single input vector element b_1, overwritten with z.
[in]strideb[hipblasStride] distance between elements of b in batch (distance between b_i and b_(i + 1))
[in,out]cdevice strided_batched pointer or host strided_batched pointer to first cosine element of Givens rotations c_1.
[in]stridec[hipblasStride] distance between elements of c in batch (distance between c_i and c_(i + 1))
[in,out]sdevice strided_batched pointer or host strided_batched pointer to sine element of Givens rotations s_1.
[in]strides[hipblasStride] distance between elements of s in batch (distance between s_i and s_(i + 1))
[in]batchCount[int] number of batches (length of arrays a, b, c, and s).

◆ hipblasCrotStridedBatched()

hipblasStatus_t hipblasCrotStridedBatched ( hipblasHandle_t  handle,
int  n,
hipblasComplex x,
int  incx,
hipblasStride  stridex,
hipblasComplex y,
int  incy,
hipblasStride  stridey,
const float *  c,
const hipblasComplex s,
int  batchCount 
)

BLAS Level 1 API.

rotStridedBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to strided batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment between elements of each x_i.
[in]stridex[hipblasStride] specifies the increment from the beginning of x_i to the beginning of x_(i+1)
[in,out]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment between elements of each y_i.
[in]stridey[hipblasStride] specifies the increment from the beginning of y_i to the beginning of y_(i+1)
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasCscal()

hipblasStatus_t hipblasCscal ( hipblasHandle_t  handle,
int  n,
const hipblasComplex alpha,
hipblasComplex x,
int  incx 
)

BLAS Level 1 API.

scal scales each element of vector x with scalar alpha.

x := alpha * x
  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]alphadevice pointer or host pointer for the scalar alpha.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasCscalBatched()

hipblasStatus_t hipblasCscalBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex alpha,
hipblasComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 1 API.

scalBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasCscalStridedBatched()

hipblasStatus_t hipblasCscalStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex alpha,
hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 1 API.

scalStridedBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i ,

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasCspr()

hipblasStatus_t hipblasCspr ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
hipblasComplex AP 
)

BLAS Level 2 API.

spr performs the matrix-vector operations

A := A + alpha*x*x**T

where alpha is a scalar, x is a vector, and A is an n by n symmetric matrix, supplied in packed form.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in AP.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing the packed version of the specified triangular portion of the symmetric matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0

◆ hipblasCsprBatched()

hipblasStatus_t hipblasCsprBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  x[],
int  incx,
hipblasComplex *const  AP[],
int  batchCount 
)

BLAS Level 2 API.

sprBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**T

where alpha is a scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each symmetric matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]batchCount[int] number of instances in the batch.

◆ hipblasCsprStridedBatched()

hipblasStatus_t hipblasCsprStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
hipblasComplex AP,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

sprStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**T

where alpha is a scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in,out]APdevice pointer storing the packed version of the specified triangular portion of each symmetric matrix A_i. Points to the first A_1. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasCsrot()

hipblasStatus_t hipblasCsrot ( hipblasHandle_t  handle,
int  n,
hipblasComplex x,
int  incx,
hipblasComplex y,
int  incy,
const float *  c,
const float *  s 
)

BLAS Level 1 API.

rot applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to vectors x and y. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment between elements of y.
[in]cdevice pointer or host pointer storing scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer storing scalar sine component of the rotation matrix.

◆ hipblasCsrotBatched()

hipblasStatus_t hipblasCsrotBatched ( hipblasHandle_t  handle,
int  n,
hipblasComplex *const  x[],
int  incx,
hipblasComplex *const  y[],
int  incy,
const float *  c,
const float *  s,
int  batchCount 
)

BLAS Level 1 API.

rotBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice array of deivce pointers storing each vector x_i.
[in]incx[int] specifies the increment between elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment between elements of each y_i.
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasCsrotStridedBatched()

hipblasStatus_t hipblasCsrotStridedBatched ( hipblasHandle_t  handle,
int  n,
hipblasComplex x,
int  incx,
hipblasStride  stridex,
hipblasComplex y,
int  incy,
hipblasStride  stridey,
const float *  c,
const float *  s,
int  batchCount 
)

BLAS Level 1 API.

rotStridedBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to strided batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment between elements of each x_i.
[in]stridex[hipblasStride] specifies the increment from the beginning of x_i to the beginning of x_(i+1)
[in,out]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment between elements of each y_i.
[in]stridey[hipblasStride] specifies the increment from the beginning of y_i to the beginning of y_(i+1)
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasCsscal()

hipblasStatus_t hipblasCsscal ( hipblasHandle_t  handle,
int  n,
const float *  alpha,
hipblasComplex x,
int  incx 
)

BLAS Level 1 API.

scal scales each element of vector x with scalar alpha.

x := alpha * x
  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]alphadevice pointer or host pointer for the scalar alpha.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasCsscalBatched()

hipblasStatus_t hipblasCsscalBatched ( hipblasHandle_t  handle,
int  n,
const float *  alpha,
hipblasComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 1 API.

scalBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasCsscalStridedBatched()

hipblasStatus_t hipblasCsscalStridedBatched ( hipblasHandle_t  handle,
int  n,
const float *  alpha,
hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 1 API.

scalStridedBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i ,

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasCswap()

hipblasStatus_t hipblasCswap ( hipblasHandle_t  handle,
int  n,
hipblasComplex x,
int  incx,
hipblasComplex y,
int  incy 
)

BLAS Level 1 API.

swap interchanges vectors x and y.

y := x; x := y
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasCswapBatched()

hipblasStatus_t hipblasCswapBatched ( hipblasHandle_t  handle,
int  n,
hipblasComplex *const  x[],
int  incx,
hipblasComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

swapBatched interchanges vectors x_i and y_i, for i = 1 , ... , batchCount

y_i := x_i; x_i := y_i
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasCswapStridedBatched()

hipblasStatus_t hipblasCswapStridedBatched ( hipblasHandle_t  handle,
int  n,
hipblasComplex x,
int  incx,
hipblasStride  stridex,
hipblasComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

swapStridedBatched interchanges vectors x_i and y_i, for i = 1 , ... , batchCount

y_i := x_i; x_i := y_i
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in,out]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in,out]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_y is of appropriate size, for a typical case this means stride_y >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch.

◆ hipblasCsymm()

hipblasStatus_t hipblasCsymm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex BP,
int  ldb,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc 
)

BLAS Level 3 API.

symm performs one of the matrix-matrix operations:

C := alpha*A*B + beta*C if side == HIPBLAS_SIDE_LEFT, C := alpha*B*A + beta*C if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B and C are m by n matrices, and A is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C := alpha*A*B + beta*C HIPBLAS_SIDE_RIGHT: C := alpha*B*A + beta*C
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix
[in]m[int] m specifies the number of rows of B and C. m >= 0.
[in]n[int] n specifies the number of columns of B and C. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A and B are not referenced.
[in]APpointer storing matrix A on the GPU. A is m by m if side == HIPBLAS_SIDE_LEFT A is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPpointer storing matrix B on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m )

◆ hipblasCsymmBatched()

hipblasStatus_t hipblasCsymmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  BP[],
int  ldb,
const hipblasComplex beta,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

symmBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]m[int] m specifies the number of rows of B_i and C_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i and C_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPdevice array of device pointers storing each matrix B_i on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C_i need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C_i. ldc >= max( 1, m )
[in]batchCount[int] number of instances in the batch.

◆ hipblasCsymmStridedBatched()

hipblasStatus_t hipblasCsymmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex BP,
int  ldb,
hipblasStride  strideB,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

symmStridedBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]m[int] m specifies the number of rows of B_i and C_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i and C_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice pointer to first matrix A_1 A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPdevice pointer to first matrix B_1 of dimension (ldb, n) on the GPU.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m )
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice pointer to first matrix C_1 of dimension (ldc, n) on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasCsymv()

hipblasStatus_t hipblasCsymv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex x,
int  incx,
const hipblasComplex beta,
hipblasComplex y,
int  incy 
)

BLAS Level 2 API.

symv performs the matrix-vector operation:

y := alpha*A*x + beta*y,

where alpha and beta are scalars, x and y are n element vectors and A should contain an upper or lower triangular n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int]
[in]alphaspecifies the scalar alpha
[in]APpointer storing matrix A on the GPU
[in]lda[int] specifies the leading dimension of A
[in]xpointer storing vector x on the GPU
[in]incx[int] specifies the increment for the elements of x
[in]betaspecifies the scalar beta
[out]ypointer storing vector y on the GPU
[in]incy[int] specifies the increment for the elements of y

◆ hipblasCsymvBatched()

hipblasStatus_t hipblasCsymvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex beta,
hipblasComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

symvBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A a should contain an upper or lower triangular symmetric matrix and the opposing triangular part of A is not referenced

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APdevice array of device pointers storing each matrix A_i
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]xdevice array of device pointers storing each vector x_i
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]betadevice pointer or host pointer to scalar beta
[out]ydevice array of device pointers storing each vector y_i
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]batchCount[int] number of instances in the batch

◆ hipblasCsymvStridedBatched()

hipblasStatus_t hipblasCsymvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex beta,
hipblasComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

symvStridedBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A a should contain an upper or lower triangular symmetric matrix and the opposing triangular part of A is not referenced

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APDevice pointer to the first matrix A_1 on the GPU
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xDevice pointer to the first vector x_1 on the GPU
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. This typically means stridex >= n * incx. stridex should be non zero.
[in]betadevice pointer or host pointer to scalar beta
[out]yDevice pointer to the first vector y_1 on the GPU
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. This typically means stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasCsyr()

hipblasStatus_t hipblasCsyr ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
hipblasComplex AP,
int  lda 
)

BLAS Level 2 API.

syr performs the matrix-vector operations

A := A + alpha*x*x**T

where alpha is a scalar, x is a vector, and A is an n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasCsyr2()

hipblasStatus_t hipblasCsyr2 ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
const hipblasComplex y,
int  incy,
hipblasComplex AP,
int  lda 
)

BLAS Level 2 API.

syr2 performs the matrix-vector operations

A := A + alpha*x*y**T + alpha*y*x**T

where alpha is a scalar, x and y are vectors, and A is an n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasCsyr2Batched()

hipblasStatus_t hipblasCsyr2Batched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  x[],
int  incx,
const hipblasComplex *const  y[],
int  incy,
hipblasComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

syr2Batched performs a batch of matrix-vector operations

A[i] := A[i] + alpha*x[i]*y[i]**T + alpha*y[i]*x[i]**T

where alpha is a scalar, x[i] and y[i] are vectors, and A[i] is a n by n symmetric matrix, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasCsyr2k()

hipblasStatus_t hipblasCsyr2k ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex BP,
int  ldb,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc 
)

BLAS Level 3 API.

syr2k performs one of the matrix-matrix operations for a symmetric rank-2k update

C := alpha*(op( A )*op( B )^T + op( B )*op( A )^T) + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^T, op( B ) = B^T,  and A and B are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A ) = A^T, op( B ) = B^T HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A) and op(B). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasCsyr2kBatched()

hipblasStatus_t hipblasCsyr2kBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  BP[],
int  ldb,
const hipblasComplex beta,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syr2kBatched performs a batch of the matrix-matrix operations for a symmetric rank-2k update

C_i := alpha*(op( A_i )*op( B_i )^T + op( B_i )*op( A_i )^T) + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCsyr2kStridedBatched()

hipblasStatus_t hipblasCsyr2kStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex BP,
int  ldb,
hipblasStride  strideB,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syr2kStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-2k update

C_i := alpha*(op( A_i )*op( B_i )^T + op( B_i )*op( A_i )^T) + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasCsyr2StridedBatched()

hipblasStatus_t hipblasCsyr2StridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
const hipblasComplex y,
int  incy,
hipblasStride  stridey,
hipblasComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

syr2StridedBatched the matrix-vector operations

A[i] := A[i] + alpha*x[i]*y[i]**T + alpha*y[i]*x[i]**T

where alpha is a scalar, x[i] and y[i] are vectors, and A[i] is a n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of each matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the pointer increment between vectors (x_i) and (x_i+1).
[in]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] specifies the pointer increment between vectors (y_i) and (y_i+1).
[in,out]APdevice pointer to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasCsyrBatched()

hipblasStatus_t hipblasCsyrBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  x[],
int  incx,
hipblasComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

syrBatched performs a batch of matrix-vector operations

A[i] := A[i] + alpha*x[i]*x[i]**T

where alpha is a scalar, x is an array of vectors, and A is an array of n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasCsyrk()

hipblasStatus_t hipblasCsyrk ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc 
)

BLAS Level 3 API.

syrk performs one of the matrix-matrix operations for a symmetric rank-k update

C := alpha*op( A )*op( A )^T + beta*C

where alpha and beta are scalars, op(A) is an n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower.

op( A ) = A, and A is n by k if transA == HIPBLAS_OP_N
op( A ) = A^T and A is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if transA = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasCsyrkBatched()

hipblasStatus_t hipblasCsyrkBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex beta,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syrkBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( A_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) is an n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^T and A_i is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCsyrkStridedBatched()

hipblasStatus_t hipblasCsyrkStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syrkStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( A_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) is an n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^T and A_i is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU. on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasCsyrkx()

hipblasStatus_t hipblasCsyrkx ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex BP,
int  ldb,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc 
)

BLAS Level 3 API.

syrkx performs one of the matrix-matrix operations for a symmetric rank-k update

C := alpha*op( A )*op( B )^T + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A )*op( B )^T will be symmetric.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^T, op( B ) = B^T,  and A and B are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A ) = A^T, op( B ) = B^T HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A) and op(B). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasCsyrkxBatched()

hipblasStatus_t hipblasCsyrkxBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  BP[],
int  ldb,
const hipblasComplex beta,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syrkxBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( B_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A_i )*op( B_i )^T will be symmetric.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCsyrkxStridedBatched()

hipblasStatus_t hipblasCsyrkxStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex BP,
int  ldb,
hipblasStride  strideB,
const hipblasComplex beta,
hipblasComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syrkxStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( B_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A_i )*op( B_i )^T will be symmetric.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasCsyrStridedBatched()

hipblasStatus_t hipblasCsyrStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasComplex alpha,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
hipblasComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

syrStridedBatched performs the matrix-vector operations

A[i] := A[i] + alpha*x[i]*x[i]**T

where alpha is a scalar, vectors, and A is an array of n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of each matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the pointer increment between vectors (x_i) and (x_i+1).
[in,out]APdevice pointer to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasCtbmv()

hipblasStatus_t hipblasCtbmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const hipblasComplex AP,
int  lda,
hipblasComplex x,
int  incx 
)

BLAS Level 2 API.

tbmv performs one of the matrix-vector operations

x := A*x      or
x := A**T*x   or
x := A**H*x,

x is a vectors and A is a banded m by m matrix (see description below).

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of A is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of A's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by A.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of the matrix A. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of the matrix A. k must satisfy k > 0 && k < lda.
[in]APdevice pointer storing banded triangular matrix A. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of A. lda must satisfy lda > k.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasCtbmvBatched()

hipblasStatus_t hipblasCtbmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const hipblasComplex *const  AP[],
int  lda,
hipblasComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tbmvBatched performs one of the matrix-vector operations

x_i := A_i*x_i      or
x_i := A_i**T*x_i   or
x_i := A_i**H*x_i,

where (A_i, x_i) is the i-th instance of the batch. x_i is a vector and A_i is an m by m matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether each matrix A_i is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of each A_i is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of each A_i's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by each A_i.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of each matrix A_i. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of each matrix A_i. k must satisfy k > 0 && k < lda.
[in]APdevice array of device pointers storing each banded triangular matrix A_i. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of each A_i. lda must satisfy lda > k.
[in,out]xdevice array of device pointer storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasCtbmvStridedBatched()

hipblasStatus_t hipblasCtbmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tbmvStridedBatched performs one of the matrix-vector operations

x_i := A_i*x_i      or
x_i := A_i**T*x_i   or
x_i := A_i**H*x_i,

where (A_i, x_i) is the i-th instance of the batch. x_i is a vector and A_i is an m by m matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether each matrix A_i is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of each A_i is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of each A_i's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by each A_i.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of each matrix A_i. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of each matrix A_i. k must satisfy k > 0 && k < lda.
[in]APdevice array to the first matrix A_i of the batch. Stores each banded triangular matrix A_i. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of each A_i. lda must satisfy lda > k.
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in,out]xdevice array to the first vector x_i of the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one x_i matrix to the next x_(i + 1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCtbsv()

hipblasStatus_t hipblasCtbsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const hipblasComplex AP,
int  lda,
hipblasComplex x,
int  incx 
)

BLAS Level 2 API.

tbsv solves

 A*x = b or A**T*x = b or A**H*x = b,

where x and b are vectors and A is a banded triangular matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular (i.e. the diagonal elements of A are not used in computations). HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of b. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of A. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of A. k >= 0.
[in]APdevice pointer storing the matrix A in banded format.
[in]lda[int] specifies the leading dimension of A. lda >= (k + 1).
[in,out]xdevice pointer storing input vector b. Overwritten by the output vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasCtbsvBatched()

hipblasStatus_t hipblasCtbsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const hipblasComplex *const  AP[],
int  lda,
hipblasComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tbsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a banded triangular matrix, for i = [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A_i*x_i = b_i HIPBLAS_OP_T: Solves A_i**T*x_i = b_i HIPBLAS_OP_C: Solves A_i**H*x_i = b_i
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of each b_i. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of each A_i. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of each A_i. k >= 0.
[in]APdevice vector of device pointers storing each matrix A_i in banded format.
[in]lda[int] specifies the leading dimension of each A_i. lda >= (k + 1).
[in,out]xdevice vector of device pointers storing each input vector b_i. Overwritten by each output vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasCtbsvStridedBatched()

hipblasStatus_t hipblasCtbsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tbsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a banded triangular matrix, for i = [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A_i*x_i = b_i HIPBLAS_OP_T: Solves A_i**T*x_i = b_i HIPBLAS_OP_C: Solves A_i**H*x_i = b_i
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of each b_i. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of each A_i. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of each A_i. k >= 0.
[in]APdevice pointer pointing to the first banded matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i. lda >= (k + 1).
[in]strideA[hipblasStride] specifies the distance between the start of one matrix (A_i) and the next (A_i+1).
[in,out]xdevice pointer pointing to the first input vector b_1. Overwritten by output vectors x.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the distance between the start of one vector (x_i) and the next (x_i+1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasCtpmv()

hipblasStatus_t hipblasCtpmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex AP,
hipblasComplex x,
int  incx 
)

BLAS Level 2 API.

tpmv performs one of the matrix-vector operations

 x = A*x or x = A**T*x,

where x is an n element vector and A is an n by n unit, or non-unit, upper or lower triangular matrix, supplied in the pack form.

The vector x is overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of A. m >= 0.
[in]APdevice pointer storing matrix A, of dimension at leat ( m * ( m + 1 ) / 2 ). Before entry with uplo = HIPBLAS_FILL_MODE_UPPER, the array A must contain the upper triangular matrix packed sequentially, column by column, so that A[0] contains a_{0,0}, A[1] and A[2] contain a_{0,1} and a_{1, 1} respectively, and so on. Before entry with uplo = HIPBLAS_FILL_MODE_LOWER, the array A must contain the lower triangular matrix packed sequentially, column by column, so that A[0] contains a_{0,0}, A[1] and A[2] contain a_{1,0} and a_{2,0} respectively, and so on. Note that when DIAG = HIPBLAS_DIAG_UNIT, the diagonal elements of A are not referenced, but are assumed to be unity.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x. incx must not be zero.

◆ hipblasCtpmvBatched()

hipblasStatus_t hipblasCtpmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex *const  AP[],
hipblasComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tpmvBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix)

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer storing pointer of matrices A_i, of dimension ( lda, m )
[in]xdevice pointer storing vectors x_i.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasCtpmvStridedBatched()

hipblasStatus_t hipblasCtpmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex AP,
hipblasStride  strideA,
hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tpmvStridedBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix) with strides specifying how to retrieve $x_i$ (resp. $A_i$) from $x_{i-1}$ (resp. $A_i$).

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer of the matrix A_0, of dimension ( lda, m )
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_{i + 1}
[in]xdevice pointer storing the vector x_0.
[in]incx[int] specifies the increment for the elements of one vector x.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_{i + 1}
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasCtpsv()

hipblasStatus_t hipblasCtpsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex AP,
hipblasComplex x,
int  incx 
)

BLAS Level 2 API.

tpsv solves

 A*x = b or A**T*x = b, or A**H*x = b,

where x and b are vectors and A is a triangular matrix stored in the packed format.

The input vector b is overwritten by the output vector x.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular (i.e. the diagonal elements of A are not used in computations). HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice pointer storing the packed version of matrix A, of dimension >= (n * (n + 1) / 2)
[in,out]xdevice pointer storing vector b on input, overwritten by x on output.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasCtpsvBatched()

hipblasStatus_t hipblasCtpsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex *const  AP[],
hipblasComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tpsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i, or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a triangular matrix stored in the packed format, for i in [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice array of device pointers storing the packed versions of each matrix A_i, of dimension >= (n * (n + 1) / 2)
[in,out]xdevice array of device pointers storing each input vector b_i, overwritten by x_i on output.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasCtpsvStridedBatched()

hipblasStatus_t hipblasCtpsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex AP,
hipblasStride  strideA,
hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tpsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i, or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a triangular matrix stored in the packed format, for i in [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice pointer pointing to the first packed matrix A_1, of dimension >= (n * (n + 1) / 2)
[in]strideA[hipblasStride] stride from the beginning of one packed matrix (AP_i) and the next (AP_i+1).
[in,out]xdevice pointer pointing to the first input vector b_1. Overwritten by each x_i on output.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the beginning of one vector (x_i) and the next (x_i+1).
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasCtrmm()

hipblasStatus_t hipblasCtrmm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasComplex BP,
int  ldb 
)

BLAS Level 3 API.

#ifndef HIPBLAS_V1 // deprecated

    hipblasStatus_t hipblasStrmm(hipblasHandle_t    handle,
                                 hipblasSideMode_t  side,
                                 hipblasFillMode_t  uplo,
                                 hipblasOperation_t transA,
                                 hipblasDiagType_t  diag,
                                 int                m,
                                 int                n,
                                 const float*       alpha,
                                 const float*       AP,
                                 int                lda,
                                 float*             BP,
                                 int                ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasStrmmOutofplace(hipblasHandle_t    handle,
                                          hipblasSideMode_t  side,
                                          hipblasFillMode_t  uplo,
                                          hipblasOperation_t transA,
                                          hipblasDiagType_t  diag,
                                          int                m,
                                          int                n,
                                          const float*       alpha,
                                          const float*       AP,
                                          int                lda,
                                          const float*       BP,
                                          int                ldb,
                                          float*             CP,
                                          int                ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasDtrmm(hipblasHandle_t    handle,
                                hipblasSideMode_t  side,
                                hipblasFillMode_t  uplo,
                                hipblasOperation_t transA,
                                hipblasDiagType_t  diag,
                                int                m,
                                int                n,
                                const double*      alpha,
                                const double*      AP,
                                int                lda,
                                double*            BP,
                                int                ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasDtrmmOutofplace(hipblasHandle_t    handle,
                                          hipblasSideMode_t  side,
                                          hipblasFillMode_t  uplo,
                                          hipblasOperation_t transA,
                                          hipblasDiagType_t  diag,
                                          int                m,
                                          int                n,
                                          const double*      alpha,
                                          const double*      AP,
                                          int                lda,
                                          const double*      BP,
                                          int                ldb,
                                          double*            CP,
                                          int                ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasCtrmm(hipblasHandle_t       handle,
                                hipblasSideMode_t     side,
                                hipblasFillMode_t     uplo,
                                hipblasOperation_t    transA,
                                hipblasDiagType_t     diag,
                                int                   m,
                                int                   n,
                                const hipblasComplex* alpha,
                                const hipblasComplex* AP,
                                int                   lda,
                                hipblasComplex*       BP,
                                int                   ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasCtrmmOutofplace(hipblasHandle_t       handle,
                                          hipblasSideMode_t     side,
                                          hipblasFillMode_t     uplo,
                                          hipblasOperation_t    transA,
                                          hipblasDiagType_t     diag,
                                          int                   m,
                                          int                   n,
                                          const hipblasComplex* alpha,
                                          const hipblasComplex* AP,
                                          int                   lda,
                                          const hipblasComplex* BP,
                                          int                   ldb,
                                          hipblasComplex*       CP,
                                          int                   ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasZtrmm(hipblasHandle_t             handle,
                                hipblasSideMode_t           side,
                                hipblasFillMode_t           uplo,
                                hipblasOperation_t          transA,
                                hipblasDiagType_t           diag,
                                int                         m,
                                int                         n,
                                const hipblasDoubleComplex* alpha,
                                const hipblasDoubleComplex* AP,
                                int                         lda,
                                hipblasDoubleComplex*       BP,
                                int                         ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasZtrmmOutofplace(hipblasHandle_t             handle,
                                          hipblasSideMode_t           side,
                                          hipblasFillMode_t           uplo,
                                          hipblasOperation_t          transA,
                                          hipblasDiagType_t           diag,
                                          int                         m,
                                          int                         n,
                                          const hipblasDoubleComplex* alpha,
                                          const hipblasDoubleComplex* AP,
                                          int                         lda,
                                          const hipblasDoubleComplex* BP,
                                          int                         ldb,
                                          hipblasDoubleComplex*       CP,
                                          int                         ldc);

#endif

The deprecated Legacy BLAS in-place trmm performs one of the matrix-matrix operations:

B := alpha*op( A )*B,   or
B := alpha*B*op( A ),

The new trmm performs one of the matrix-matrix operations:

C := alpha*op( A )*B,   or
C := alpha*B*op( A ),

The in-place functionality is still available in the new trmmm by setting pointer C equal to pointer B, and ldc equal to ldb.

alpha  is a scalar,  B  is an m by n matrix, C  is an m by n matrix,  A  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A )  is one  of

op( A ) = A     or
op( A ) = A^T   or
op( A ) = A^H.

When uplo == rocblas_fill_upper the  leading  k by k
upper triangular part of the array  A must contain the upper
triangular matrix and the strictly lower triangular part of
A is not referenced. Here k is m when side == rocblas_side_left
and is n when side == rocblas_side_right.

When uplo == rocblas_fill_lower the  leading  k by k
lower triangular part of the array  A must contain the lower
triangular matrix  and the strictly upper triangular part of
A is not referenced. Here k is m when  side == rocblas_side_left
and is n when side == rocblas_side_right.

Note that when  diag == rocblas_diagonal_unit  the diagonal elements of
A  are not referenced either,  but are assumed to be  unity.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A) multiplies B from the left or right as follows: HIPBLAS_SIDE_LEFT: B := alpha*op( A )*B. HIPBLAS_SIDE_RIGHT: B := alpha*B*op( A ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A is unit triangular as follows: HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B. m >= 0.
[in]n[int] n specifies the number of columns of B. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and B need not be set before entry.
[in]APDevice pointer to matrix A on the GPU. A has dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPDevice pointer to the first matrix B_0 on the GPU. On entry, the leading m by n part of the array B must contain the matrix B, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m ).

◆ hipblasCtrmmBatched()

hipblasStatus_t hipblasCtrmmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
hipblasComplex *const  BP[],
int  ldb,
int  batchCount 
)

BLAS Level 3 API.

The hipBLAS trmm_batched API is from Legacy BLAS and it supports only in-place functionality. It is deprecated and it will be replaced with an API that supports both in-place and out-of-place functionality. The new API is available in hipBLAS versions 1.x.x and later. To get the new API compile with the directive -DHIPBLAS_V1.

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasStrmmBatched(hipblasHandle_t    handle,
                                        hipblasSideMode_t  side,
                                        hipblasFillMode_t  uplo,
                                        hipblasOperation_t transA,
                                        hipblasDiagType_t  diag,
                                        int                m,
                                        int                n,
                                        const float*       alpha,
                                        const float* const AP[],
                                        int                lda,
                                        float* const       BP[],
                                        int                ldb,
                                        int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasStrmmBatchedOutofplace(
                                        hipblasHandle_t    handle,
                                        hipblasSideMode_t  side,
                                        hipblasFillMode_t  uplo,
                                        hipblasOperation_t transA,
                                        hipblasDiagType_t  diag,
                                        int                m,
                                        int                n,
                                        const float*       alpha,
                                        const float* const AP[],
                                        int                lda,
                                        const float* const BP[],
                                        int                ldb,
                                        float* const       CP[],
                                        int                ldc,
                                        int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasDtrmmBatched(hipblasHandle_t     handle,
                                        hipblasSideMode_t   side,
                                        hipblasFillMode_t   uplo,
                                        hipblasOperation_t  transA,
                                        hipblasDiagType_t   diag,
                                        int                 m,
                                        int                 n,
                                        const double*       alpha,
                                        const double* const AP[],
                                        int                 lda,
                                        double* const       BP[],
                                        int                 ldb,
                                        int                 batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasDtrmmBatchedOutofplace(
                                        hipblasHandle_t     handle,
                                        hipblasSideMode_t   side,
                                        hipblasFillMode_t   uplo,
                                        hipblasOperation_t  transA,
                                        hipblasDiagType_t   diag,
                                        int                 m,
                                        int                 n,
                                        const double*       alpha,
                                        const double* const AP[],
                                        int                 lda,
                                        const double* const BP[],
                                        int                 ldb,
                                        double* const       CP[],
                                        int                 ldc,
                                        int                 batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasCtrmmBatched(hipblasHandle_t             handle,
                                        hipblasSideMode_t           side,
                                        hipblasFillMode_t           uplo,
                                        hipblasOperation_t          transA,
                                        hipblasDiagType_t           diag,
                                        int                         m,
                                        int                         n,
                                        const hipblasComplex*       alpha,
                                        const hipblasComplex* const AP[],
                                        int                         lda,
                                        hipblasComplex* const       BP[],
                                        int                         ldb,
                                        int                         batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasCtrmmBatchedOutofplace(
                                        hipblasHandle_t             handle,
                                        hipblasSideMode_t           side,
                                        hipblasFillMode_t           uplo,
                                        hipblasOperation_t          transA,
                                        hipblasDiagType_t           diag,
                                        int                         m,
                                        int                         n,
                                        const hipblasComplex*       alpha,
                                        const hipblasComplex* const AP[],
                                        int                         lda,
                                        const hipblasComplex* const BP[],
                                        int                         ldb,
                                        hipblasComplex* const       CP[],
                                        int                         ldc,
                                        int batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasZtrmmBatched(hipblasHandle_t                   handle,
                                        hipblasSideMode_t                 side,
                                        hipblasFillMode_t                 uplo,
                                        hipblasOperation_t                transA,
                                        hipblasDiagType_t                 diag,
                                        int                               m,
                                        int                               n,
                                        const hipblasDoubleComplex*       alpha,
                                        const hipblasDoubleComplex* const AP[],
                                        int                               lda,
                                        hipblasDoubleComplex* const       BP[],
                                        int                               ldb,
                                        int                               batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasZtrmmBatchedOutofplace(
                                        hipblasHandle_t                   handle,
                                        hipblasSideMode_t                 side,
                                        hipblasFillMode_t                 uplo,
                                        hipblasOperation_t                transA,
                                        hipblasDiagType_t                 diag,
                                        int                               m,
                                        int                               n,
                                        const hipblasDoubleComplex*       alpha,
                                        const hipblasDoubleComplex* const AP[],
                                        int                               lda,
                                        const hipblasDoubleComplex* const BP[],
                                        int                               ldb,
                                        hipblasDoubleComplex* const       CP[],
                                        int                               ldc,
                                        int                               batchCount);

#endif

The deprecated Legacy BLAS in-place trmm_batched performs one of the batched matrix-matrix operations:

B_i := alpha*op( A_i )*B_i,   or
B_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The new trmm_batched performs one of the matrix-matrix operations:

C_i := alpha*op( A_i )*B_i,   or
C_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The in-place functionality is still available in the new trmmm_batched by setting pointer C equal to pointer B and ldc equal to ldb.

alpha  is a scalar,  B_i  is an m by n matrix, C_i  is an m by n matrix,  A_i  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A_i )  is one  of

op( A_i ) = A_i   or   op( A_i ) = A_i^T   or   op( A_i ) = A_i^H.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A_i) multiplies B_i from the left or right as follows: HIPBLAS_SIDE_LEFT: B_i := alpha*op( A_i )*B_i. HIPBLAS_SIDE_RIGHT: B_i := alpha*B_i*op( A_i ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A_i) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A_i) = A_i. HIPBLAS_OP_T: op(A_i) = A_i^T. HIPBLAS_OP_C: op(A_i) = A_i^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A_i is unit triangular as follows: HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i is not referenced and B_i need not be set before entry.
[in]APDevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A_i are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice array of device pointers storing each matrix B_i on the GPU. On entry, the leading m by n part of the array B_i must contain the matrix B_i, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m ).
[in]batchCount[int] number of instances i in the batch.

◆ hipblasCtrmmBatchedOutofplace()

hipblasStatus_t hipblasCtrmmBatchedOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
const hipblasComplex *const  BP[],
int  ldb,
hipblasComplex *const  CP[],
int  ldc,
int  batchCount 
)

◆ hipblasCtrmmOutofplace()

hipblasStatus_t hipblasCtrmmOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
const hipblasComplex BP,
int  ldb,
hipblasComplex CP,
int  ldc 
)

◆ hipblasCtrmmStridedBatched()

hipblasStatus_t hipblasCtrmmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
hipblasComplex BP,
int  ldb,
hipblasStride  strideB,
int  batchCount 
)

BLAS Level 3 API.

The hipBLAS trmm_strided_batched API is from Legacy BLAS and it supports only in-place functionality. It is deprecated and it will be replaced with an API that supports both in-place and out-of-place functionality. The new API is available in hipBLAS versions 1.x.x and later. To get the new API compile with the directive -DHIPBLAS_V1.

#ifndef HIPBLAS_V1  //  deprecated
       hipblasStatus_t hipblasStrmmStridedBatched(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const float*       alpha,
                           const float*       AP,
                           int                lda,
                           hipblasStride      strideA,
                           float*             BP,
                           int                ldb,
                           hipblasStride      strideB,
                           int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasStrmmStridedBatchedOutofplace(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const float*       alpha,
                           const float*       AP,
                           int                lda,
                           hipblasStride      strideA,
                           const float*       BP,
                           int                ldb,
                           hipblasStride      strideB,
                           float*             CP,
                           int                ldc,
                           hipblasStride      strideC,
                           int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated
       hipblasStatus_t hipblasDtrmmStridedBatched(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const double*      alpha,
                           const double*      AP,
                           int                lda,
                           hipblasStride      strideA,
                           double*            BP,
                           int                ldb,
                           hipblasStride      strideB,
                           int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasDtrmmStridedBatchedOutofplace(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const double*      alpha,
                           const double*      AP,
                           int                lda,
                           hipblasStride      strideA,
                           const double*      BP,
                           int                ldb,
                           hipblasStride      strideB,
                           double*            CP,
                           int                ldc,
                           hipblasStride      strideC,
                           int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

       hipblasStatus_t hipblasCtrmmStridedBatched(
                           hipblasHandle_t       handle,
                           hipblasSideMode_t     side,
                           hipblasFillMode_t     uplo,
                           hipblasOperation_t    transA,
                           hipblasDiagType_t     diag,
                           int                   m,
                           int                   n,
                           const hipblasComplex* alpha,
                           const hipblasComplex* AP,
                           int                   lda,
                           hipblasStride         strideA,
                           hipblasComplex*       BP,
                           int                   ldb,
                           hipblasStride         strideB,
                           int                   batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasCtrmmStridedBatchedOutofplace(
                           hipblasHandle_t       handle,
                           hipblasSideMode_t     side,
                           hipblasFillMode_t     uplo,
                           hipblasOperation_t    transA,
                           hipblasDiagType_t     diag,
                           int                   m,
                           int                   n,
                           const hipblasComplex* alpha,
                           const hipblasComplex* AP,
                           int                   lda,
                           hipblasStride         strideA,
                           const hipblasComplex* BP,
                           int                   ldb,
                           hipblasStride         strideB,
                           hipblasComplex*       CP,
                           int                   ldc,
                           hipblasStride         strideC,
                           int batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasZtrmmStridedBatched(hipblasHandle_t             handle,
                                               hipblasSideMode_t           side,
                                               hipblasFillMode_t           uplo,
                                               hipblasOperation_t          transA,
                                               hipblasDiagType_t           diag,
                                               int                         m,
                                               int                         n,
                                               const hipblasDoubleComplex* alpha,
                                               const hipblasDoubleComplex* AP,
                                               int                         lda,
                                               hipblasStride               strideA,
                                               hipblasDoubleComplex*       BP,
                                               int                         ldb,
                                               hipblasStride               strideB,
                                               int                         batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasZtrmmStridedBatchedOutofplace(
                                                hipblasHandle_t             handle,
                                                hipblasSideMode_t           side,
                                                hipblasFillMode_t           uplo,
                                                hipblasOperation_t          transA,
                                                hipblasDiagType_t           diag,
                                                int                         m,
                                                int                         n,
                                                const hipblasDoubleComplex* alpha,
                                                const hipblasDoubleComplex* AP,
                                                int                         lda,
                                                hipblasStride               strideA,
                                                const hipblasDoubleComplex* BP,
                                                int                         ldb,
                                                hipblasStride               strideB,
                                                hipblasDoubleComplex*       BC,
                                                int                         ldc,
                                                hipblasStride               strideC,
                                                int                         batchCount);

#endif

The deprecated Legacy BLAS in-place trmm_strided_batched performs one of the strided_batched matrix-matrix operations:

B_i := alpha*op( A_i )*B_i,   or
B_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The new trmm_batched performs one of the matrix-matrix operations:

C_i := alpha*op( A_i )*B_i,   or
C_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The in-place functionality is still available in the new trmmm_batched by setting pointer C equal to pointer B, setting ldc equal to ldb, and setting stride_C equal to stride_B.

alpha  is a scalar,  B_i  is an m by n matrix, C_i  is an m by n matrix,  A_i  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A_i )  is one  of

op( A_i ) = A_i   or
op( A_i ) = A_i^T   or
op( A_i ) = A_i^H.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A_i) multiplies B_i from the left or right as follows: HIPBLAS_SIDE_LEFT: B_i := alpha*op( A_i )*B_i. HIPBLAS_SIDE_RIGHT: B_i := alpha*B_i*op( A_i ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A_i) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A_i) = A_i. HIPBLAS_OP_T: op(A_i) = A_i^T. HIPBLAS_OP_C: op(A_i) = A_i^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A_i is unit triangular as follows: HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i is not referenced and B_i need not be set before entry.
[in]APDevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A_i are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in,out]BPDevice pointer to the first matrix B_0 on the GPU. On entry, the leading m by n part of the array B_i must contain the matrix B_i, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasCtrmmStridedBatchedOutofplace()

hipblasStatus_t hipblasCtrmmStridedBatchedOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasComplex BP,
int  ldb,
hipblasStride  strideB,
hipblasComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

◆ hipblasCtrmv()

hipblasStatus_t hipblasCtrmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex AP,
int  lda,
hipblasComplex x,
int  incx 
)

BLAS Level 2 API.

trmv performs one of the matrix-vector operations

 x = A*x or x = A**T*x,

where x is an n element vector and A is an n by n unit, or non-unit, upper or lower triangular matrix.

The vector x is overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of A. m >= 0.
[in]APdevice pointer storing matrix A, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A. lda = max( 1, m ).
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasCtrmvBatched()

hipblasStatus_t hipblasCtrmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex *const  AP[],
int  lda,
hipblasComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

trmvBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix)

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer storing pointer of matrices A_i, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A_i. lda >= max( 1, m ).
[in]xdevice pointer storing vectors x_i.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasCtrmvStridedBatched()

hipblasStatus_t hipblasCtrmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

trmvStridedBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix) with strides specifying how to retrieve $x_i$ (resp. $A_i$) from $x_{i-1}$ (resp. $A_i$).

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer of the matrix A_0, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A_i. lda >= max( 1, m ).
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_{i + 1}
[in]xdevice pointer storing the vector x_0.
[in]incx[int] specifies the increment for the elements of one vector x.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_{i + 1}
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasCtrsm()

hipblasStatus_t hipblasCtrsm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasComplex BP,
int  ldb 
)

BLAS Level 3 API.

trsm solves

op(A)*X = alpha*B or  X*op(A) = alpha*B,

where alpha is a scalar, X and B are m by n matrices, A is triangular matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

The matrix X is overwritten on B.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time.

(where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B. m >= 0.
[in]n[int] n specifies the number of columns of B. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice pointer storing matrix A. of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice pointer storing matrix B.
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m ).

◆ hipblasCtrsmBatched()

hipblasStatus_t hipblasCtrsmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex *const  AP[],
int  lda,
hipblasComplex *const  BP[],
int  ldb,
int  batchCount 
)

BLAS Level 3 API.

trsmBatched performs the following batched operation:

op(A_i)*X_i = alpha*B_i or  X_i*op(A_i) = alpha*B_i, for i = 1, ..., batchCount.

where alpha is a scalar, X and B are batched m by n matrices, A is triangular batched matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

Each matrix X_i is overwritten on B_i for i = 1, ..., batchCount.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time. (where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each B_i. m >= 0.
[in]n[int] n specifies the number of columns of each B_i. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Matricies are of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of each A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice array of device pointers storing each matrix B_i on the GPU.
[in]ldb[int] ldb specifies the first dimension of each B_i. ldb >= max( 1, m ).
[in]batchCount[int] number of trsm operatons in the batch.

◆ hipblasCtrsmStridedBatched()

hipblasStatus_t hipblasCtrsmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasComplex alpha,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
hipblasComplex BP,
int  ldb,
hipblasStride  strideB,
int  batchCount 
)

BLAS Level 3 API.

trsmSridedBatched performs the following strided batched operation:

op(A_i)*X_i = alpha*B_i or  X_i*op(A_i) = alpha*B_i, for i = 1, ..., batchCount.

where alpha is a scalar, X and B are strided batched m by n matrices, A is triangular strided batched matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

Each matrix X_i is overwritten on B_i for i = 1, ..., batchCount.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time. (where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each B_i. m >= 0.
[in]n[int] n specifies the number of columns of each B_i. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice pointer pointing to the first matrix A_1. of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of each A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in,out]BPdevice pointer pointing to the first matrix B_1.
[in]ldb[int] ldb specifies the first dimension of each B_i. ldb >= max( 1, m ).
[in]strideB[hipblasStride] stride from the start of one B_i matrix to the next B_(i + 1).
[in]batchCount[int] number of trsm operatons in the batch.

◆ hipblasCtrsv()

hipblasStatus_t hipblasCtrsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex AP,
int  lda,
hipblasComplex x,
int  incx 
)

BLAS Level 2 API.

trsv solves

 A*x = b or A**T*x = b,

where x and b are vectors and A is a triangular matrix.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice pointer storing matrix A, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A. lda = max( 1, m ).
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasCtrsvBatched()

hipblasStatus_t hipblasCtrsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex *const  AP[],
int  lda,
hipblasComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

trsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i,

where (A_i, x_i, b_i) is the i-th instance of the batch. x_i and b_i are vectors and A_i is an m by m triangular matrix.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i. lda = max(1, m)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of x.
[in]batchCount[int] number of instances in the batch

◆ hipblasCtrsvStridedBatched()

hipblasStatus_t hipblasCtrsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

trsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i,

where (A_i, x_i, b_i) is the i-th instance of the batch. x_i and b_i are vectors and A_i is an m by m triangular matrix, for i = 1, ..., batchCount.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice pointer to the first matrix (A_1) in the batch, of dimension ( lda, m )
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1)
[in]lda[int] specifies the leading dimension of each A_i. lda = max( 1, m ).
[in,out]xdevice pointer to the first vector (x_1) in the batch.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_(i + 1)
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasCtrtri()

hipblasStatus_t hipblasCtrtri ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const hipblasComplex AP,
int  lda,
hipblasComplex invA,
int  ldinvA 
)

BLAS Level 3 API.

trtri compute the inverse of a matrix A, namely, invA

and write the result into invA;
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int] size of matrix A and invA
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[out]invAdevice pointer storing matrix invA.
[in]ldinvA[int] specifies the leading dimension of invA.

◆ hipblasCtrtriBatched()

hipblasStatus_t hipblasCtrtriBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const hipblasComplex *const  AP[],
int  lda,
hipblasComplex invA[],
int  ldinvA,
int  batchCount 
)

BLAS Level 3 API.

trtriBatched compute the inverse of A_i and write into invA_i where A_i and invA_i are the i-th matrices in the batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER'
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int]
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[out]invAdevice array of device pointers storing the inverse of each matrix A_i. Partial inplace operation is supported, see below. If UPLO = 'U', the leading N-by-N upper triangular part of the invA will store the inverse of the upper triangular matrix, and the strictly lower triangular part of invA is cleared. If UPLO = 'L', the leading N-by-N lower triangular part of the invA will store the inverse of the lower triangular matrix, and the strictly upper triangular part of invA is cleared.
[in]ldinvA[int] specifies the leading dimension of each invA_i.
[in]batchCount[int] numbers of matrices in the batch

◆ hipblasCtrtriStridedBatched()

hipblasStatus_t hipblasCtrtriStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const hipblasComplex AP,
int  lda,
hipblasStride  strideA,
hipblasComplex invA,
int  ldinvA,
hipblasStride  stride_invA,
int  batchCount 
)

BLAS Level 3 API.

trtriStridedBatched compute the inverse of A_i and write into invA_i where A_i and invA_i are the i-th matrices in the batch, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER'
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int]
[in]APdevice pointer pointing to address of first matrix A_1.
[in]lda[int] specifies the leading dimension of each A.
[in]strideA[hipblasStride] "batch stride a": stride from the start of one A_i matrix to the next A_(i + 1).
[out]invAdevice pointer storing the inverses of each matrix A_i. Partial inplace operation is supported, see below. If UPLO = 'U', the leading N-by-N upper triangular part of the invA will store the inverse of the upper triangular matrix, and the strictly lower triangular part of invA is cleared. If UPLO = 'L', the leading N-by-N lower triangular part of the invA will store the inverse of the lower triangular matrix, and the strictly upper triangular part of invA is cleared.
[in]ldinvA[int] specifies the leading dimension of each invA_i.
[in]stride_invA[hipblasStride] "batch stride invA": stride from the start of one invA_i matrix to the next invA_(i + 1).
[in]batchCount[int] numbers of matrices in the batch

◆ hipblasDasum()

hipblasStatus_t hipblasDasum ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
double *  result 
)

BLAS Level 1 API.

asum computes the sum of the magnitudes of elements of a real vector x, or the sum of magnitudes of the real and imaginary parts of elements if x is a complex vector.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x. incx must be > 0.
[in,out]resultdevice pointer or host pointer to store the asum product. return is 0.0 if n <= 0.

◆ hipblasDasumBatched()

hipblasStatus_t hipblasDasumBatched ( hipblasHandle_t  handle,
int  n,
const double *const  x[],
int  incx,
int  batchCount,
double *  result 
)

BLAS Level 1 API.

asumBatched computes the sum of the magnitudes of the elements in a batch of real vectors x_i, or the sum of magnitudes of the real and imaginary parts of elements if x_i is a complex vector, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch.
[out]resultdevice array or host array of batchCount size for results. return is 0.0 if n, incx<=0.

◆ hipblasDasumStridedBatched()

hipblasStatus_t hipblasDasumStridedBatched ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
hipblasStride  stridex,
int  batchCount,
double *  result 
)

BLAS Level 1 API.

asumStridedBatched computes the sum of the magnitudes of elements of a real vectors x_i, or the sum of magnitudes of the real and imaginary parts of elements if x_i is a complex vector, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array for storing contiguous batchCount results. return is 0.0 if n, incx<=0.

◆ hipblasDaxpy()

hipblasStatus_t hipblasDaxpy ( hipblasHandle_t  handle,
int  n,
const double *  alpha,
const double *  x,
int  incx,
double *  y,
int  incy 
)

BLAS Level 1 API.

axpy computes constant alpha multiplied by vector x, plus vector y

y := alpha * x + y
  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]alphadevice pointer or host pointer to specify the scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[out]ydevice pointer storing vector y.
[in,out]incy[int] specifies the increment for the elements of y.

◆ hipblasDaxpyBatched()

hipblasStatus_t hipblasDaxpyBatched ( hipblasHandle_t  handle,
int  n,
const double *  alpha,
const double *const  x[],
int  incx,
double *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

axpyBatched compute y := alpha * x + y over a set of batched vectors.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]alphaspecifies the scalar alpha.
[in]xpointer storing vector x on the GPU.
[in]incx[int] specifies the increment for the elements of x.
[out]ypointer storing vector y on the GPU.
[in,out]incy[int] specifies the increment for the elements of y.
[in]batchCount[int] number of instances in the batch

◆ hipblasDaxpyStridedBatched()

hipblasStatus_t hipblasDaxpyStridedBatched ( hipblasHandle_t  handle,
int  n,
const double *  alpha,
const double *  x,
int  incx,
hipblasStride  stridex,
double *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

axpyStridedBatched compute y := alpha * x + y over a set of strided batched vectors.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int]
[in]alphaspecifies the scalar alpha.
[in]xpointer storing vector x on the GPU.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] specifies the increment between vectors of x.
[out]ypointer storing vector y on the GPU.
[in,out]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] specifies the increment between vectors of y.
[in]batchCount[int] number of instances in the batch

◆ hipblasDcopy()

hipblasStatus_t hipblasDcopy ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
double *  y,
int  incy 
)

BLAS Level 1 API.

copy copies each element x[i] into y[i], for i = 1 , ... , n

y := x,
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x to be copied to y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasDcopyBatched()

hipblasStatus_t hipblasDcopyBatched ( hipblasHandle_t  handle,
int  n,
const double *const  x[],
int  incx,
double *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

copyBatched copies each element x_i[j] into y_i[j], for j = 1 , ... , n; i = 1 , ... , batchCount

y_i := x_i,

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i to be copied to y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasDcopyStridedBatched()

hipblasStatus_t hipblasDcopyStridedBatched ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
hipblasStride  stridex,
double *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

copyStridedBatched copies each element x_i[j] into y_i[j], for j = 1 , ... , n; i = 1 , ... , batchCount

y_i := x_i,

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i to be copied to y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increments for the elements of vectors x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of vectors y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stride_y, however the user should take care to ensure that stride_y is of appropriate size, for a typical case this means stride_y >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasDdgmm()

hipblasStatus_t hipblasDdgmm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const double *  AP,
int  lda,
const double *  x,
int  incx,
double *  CP,
int  ldc 
)

BLAS Level 3 API.

dgmm performs one of the matrix-matrix operations

C = A * diag(x) if side == HIPBLAS_SIDE_RIGHT
C = diag(x) * A if side == HIPBLAS_SIDE_LEFT

where C and A are m by n dimensional matrices. diag( x ) is a diagonal matrix and x is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between values of x
[in,out]CPdevice pointer storing matrix C.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasDdgmmBatched()

hipblasStatus_t hipblasDdgmmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const double *const  AP[],
int  lda,
const double *const  x[],
int  incx,
double *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

dgmmBatched performs one of the batched matrix-matrix operations

C_i = A_i * diag(x_i) for i = 0, 1, ... batchCount-1 if side == HIPBLAS_SIDE_RIGHT
C_i = diag(x_i) * A_i for i = 0, 1, ... batchCount-1 if side == HIPBLAS_SIDE_LEFT

where C_i and A_i are m by n dimensional matrices. diag(x_i) is a diagonal matrix and x_i is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, n )
[in]lda[int] specifies the leading dimension of A_i.
[in]xdevice array of device pointers storing each vector x_i on the GPU. Each x_i is of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT
[in]incx[int] specifies the increment between values of x_i
[in,out]CPdevice array of device pointers storing each matrix C_i on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasDdgmmStridedBatched()

hipblasStatus_t hipblasDdgmmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const double *  AP,
int  lda,
hipblasStride  strideA,
const double *  x,
int  incx,
hipblasStride  stridex,
double *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

dgmmStridedBatched performs one of the batched matrix-matrix operations

C_i = A_i * diag(x_i)   if side == HIPBLAS_SIDE_RIGHT   for i = 0, 1, ... batchCount-1
C_i = diag(x_i) * A_i   if side == HIPBLAS_SIDE_LEFT    for i = 0, 1, ... batchCount-1

where C_i and A_i are m by n dimensional matrices. diag(x_i) is a diagonal matrix and x_i is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, n )
[in]lda[int] specifies the leading dimension of A.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xpointer to the first vector x_0 on the GPU. Each x_i is of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT
[in]incx[int] specifies the increment between values of x
[in]stridex[hipblasStride] stride from the start of one vector(x_i) and the next one (x_i+1)
[in,out]CPdevice pointer to the first matrix C_0 on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasDdot()

hipblasStatus_t hipblasDdot ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
const double *  y,
int  incy,
double *  result 
)

BLAS Level 1 API.

dot(u) performs the dot product of vectors x and y

result = x * y;

dotc performs the dot product of the conjugate of complex vector x and complex vector y

result = conjugate (x) * y;
  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the dot product. return is 0.0 if n <= 0.

◆ hipblasDdotBatched()

hipblasStatus_t hipblasDdotBatched ( hipblasHandle_t  handle,
int  n,
const double *const  x[],
int  incx,
const double *const  y[],
int  incy,
int  batchCount,
double *  result 
)

BLAS Level 1 API.

dotBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasDdotStridedBatched()

hipblasStatus_t hipblasDdotStridedBatched ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
hipblasStride  stridex,
const double *  y,
int  incy,
hipblasStride  stridey,
int  batchCount,
double *  result 
)

BLAS Level 1 API.

dotStridedBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcStridedBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasDestroy()

hipblasStatus_t hipblasDestroy ( hipblasHandle_t  handle)

Destroys the library context created using hipblasCreate()

◆ hipblasDgbmv()

hipblasStatus_t hipblasDgbmv ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const double *  alpha,
const double *  AP,
int  lda,
const double *  x,
int  incx,
const double *  beta,
double *  y,
int  incy 
)

BLAS Level 2 API.

gbmv performs one of the matrix-vector operations

y := alpha*A*x    + beta*y,   or
y := alpha*A**T*x + beta*y,   or
y := alpha*A**H*x + beta*y,

where alpha and beta are scalars, x and y are vectors and A is an m by n banded matrix with kl sub-diagonals and ku super-diagonals.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]kl[int] number of sub-diagonals of A
[in]ku[int] number of super-diagonals of A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing banded matrix A. Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of A. Must be >= (kl + ku + 1)
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasDgbmvBatched()

hipblasStatus_t hipblasDgbmvBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const double *  alpha,
const double *const  AP[],
int  lda,
const double *const  x[],
int  incx,
const double *  beta,
double *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

gbmvBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n banded matrix with kl sub-diagonals and ku super-diagonals, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of each matrix A_i
[in]n[int] number of columns of each matrix A_i
[in]kl[int] number of sub-diagonals of each A_i
[in]ku[int] number of super-diagonals of each A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each banded matrix A_i. Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of each A_i. Must be >= (kl + ku + 1)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasDgbmvStridedBatched()

hipblasStatus_t hipblasDgbmvStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
const double *  x,
int  incx,
hipblasStride  stridex,
const double *  beta,
double *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

gbmvStridedBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n banded matrix with kl sub-diagonals and ku super-diagonals, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]kl[int] number of sub-diagonals of A
[in]ku[int] number of super-diagonals of A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer to first banded matrix (A_1). Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of A. Must be >= (kl + ku + 1)
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xdevice pointer to first vector (x_1).
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer to first vector (y_1).
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (x_i+1)
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasDgeam()

hipblasStatus_t hipblasDgeam ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
const double *  beta,
const double *  BP,
int  ldb,
double *  CP,
int  ldc 
)

BLAS Level 3 API.

geam performs one of the matrix-matrix operations

C = alpha*op( A ) + beta*op( B ),

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are matrices, with op( A ) an m by n matrix, op( B ) an m by n matrix, and C an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPdevice pointer storing matrix B.
[in]ldb[int] specifies the leading dimension of B.
[in,out]CPdevice pointer storing matrix C.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasDgeamBatched()

hipblasStatus_t hipblasDgeamBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const double *  alpha,
const double *const  AP[],
int  lda,
const double *  beta,
const double *const  BP[],
int  ldb,
double *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

geamBatched performs one of the batched matrix-matrix operations

C_i = alpha*op( A_i ) + beta*op( B_i )  for i = 0, 1, ... batchCount - 1

where alpha and beta are scalars, and op(A_i), op(B_i) and C_i are m by n matrices and op( X ) is one of

op( X ) = X      or
op( X ) = X**T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, k ), where k is m when transA == HIPBLAS_OP_N and is n when transA == HIPBLAS_OP_T.
[in]lda[int] specifies the leading dimension of A.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPdevice array of device pointers storing each matrix B_i on the GPU. Each B_i is of dimension ( ldb, k ), where k is m when transB == HIPBLAS_OP_N and is n when transB == HIPBLAS_OP_T.
[in]ldb[int] specifies the leading dimension of B.
[in,out]CPdevice array of device pointers storing each matrix C_i on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]batchCount[int] number of instances i in the batch.

◆ hipblasDgeamStridedBatched()

hipblasStatus_t hipblasDgeamStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
const double *  beta,
const double *  BP,
int  ldb,
hipblasStride  strideB,
double *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

geamStridedBatched performs one of the batched matrix-matrix operations

C_i = alpha*op( A_i ) + beta*op( B_i )  for i = 0, 1, ... batchCount - 1

where alpha and beta are scalars, and op(A_i), op(B_i) and C_i are m by n matrices and op( X ) is one of

op( X ) = X      or
op( X ) = X**T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, k ), where k is m when transA == HIPBLAS_OP_N and is n when transA == HIPBLAS_OP_T.
[in]lda[int] specifies the leading dimension of A.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPpointer to the first matrix B_0 on the GPU. Each B_i is of dimension ( ldb, k ), where k is m when transB == HIPBLAS_OP_N and is n when transB == HIPBLAS_OP_T.
[in]ldb[int] specifies the leading dimension of B.
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in,out]CPpointer to the first matrix C_0 on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasDgels()

hipblasStatus_t hipblasDgels ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
double *  A,
const int  lda,
double *  B,
const int  ldb,
int *  info,
int *  deviceInfo 
)

GELS solves an overdetermined (or underdetermined) linear system defined by an m-by-n matrix A, and a corresponding matrix B, using the QR factorization computed by GEQRF (or the LQ factorization computed by "GELQF").

Depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A X = B & \: \text{not transposed, or}\\ A' X = B & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X is found by minimizing

\[ || B - A X || \quad \text{(or} \: || B - A' X ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X is chosen such that \(|| X ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : currently unsupported
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of matrix A.
[in]nint. n >= 0.
The number of columns of matrix A.
[in]nrhsint. nrhs >= 0.
The number of columns of matrices B and X; i.e., the columns on the right hand side.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the matrix A. On exit, the QR (or LQ) factorization of A as returned by "GEQRF" (or "GELQF").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrix A.
[in,out]Bpointer to type. Array on the GPU of dimension ldb*nrhs.
On entry, the matrix B. On exit, when info = 0, B is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrix B.
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int on the GPU.
If info = 0, successful exit. If info = i > 0, the solution could not be computed because input matrix A is rank deficient; the i-th diagonal element of its triangular factor is zero.

◆ hipblasDgelsBatched()

hipblasStatus_t hipblasDgelsBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
double *const  A[],
const int  lda,
double *const  B[],
const int  ldb,
int *  info,
int *  deviceInfo,
const int  batchCount 
)

gelsBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_BATCHED" (or the LQ factorizations computed by "GELQF_BATCHED").

For each instance in the batch, depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A_j X_j = B_j & \: \text{not transposed, or}\\ A_j' X_j = B_j & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X_j is found by minimizing

\[ || B_j - A_j X_j || \quad \text{(or} \: || B_j - A_j' X_j ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X_j is chosen such that \(|| X_j ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z Note that cuBLAS backend supports only the non-transpose operation and only solves over-determined systems (m >= n).
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of all matrices A_j in the batch.
[in]nint. n >= 0.
The number of columns of all matrices A_j in the batch.
[in]nrhsint. nrhs >= 0.
The number of columns of all matrices B_j and X_j in the batch; i.e., the columns on the right hand side.
[in,out]Aarray of pointer to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the matrices A_j. On exit, the QR (or LQ) factorizations of A_j as returned by "GEQRF_BATCHED" (or "GELQF_BATCHED").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_j.
[in,out]Barray of pointer to type. Each pointer points to an array on the GPU of dimension ldb*nrhs.
On entry, the matrices B_j. On exit, when info[j] = 0, B_j is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrices B_j.
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int. Array of batchCount integers on the GPU.
If deviceInfo[j] = 0, successful exit for solution of A_j. If deviceInfo[j] = i > 0, the solution of A_j could not be computed because input matrix A_j is rank deficient; the i-th diagonal element of its triangular factor is zero.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasDgelsStridedBatched()

hipblasStatus_t hipblasDgelsStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
double *  A,
const int  lda,
const hipblasStride  strideA,
double *  B,
const int  ldb,
const hipblasStride  strideB,
int *  info,
int *  deviceInfo,
const int  batch_count 
)

gelsStridedBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_STRIDED_BATCHED" (or the LQ factorizations computed by "GELQF_STRIDED_BATCHED").

For each instance in the batch, depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A_j X_j = B_j & \: \text{not transposed, or}\\ A_j' X_j = B_j & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X_j is found by minimizing

\[ || B_j - A_j X_j || \quad \text{(or} \: || B_j - A_j' X_j ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X_j is chosen such that \(|| X_j ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : currently unsupported
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of all matrices A_j in the batch.
[in]nint. n >= 0.
The number of columns of all matrices A_j in the batch.
[in]nrhsint. nrhs >= 0.
The number of columns of all matrices B_j and X_j in the batch; i.e., the columns on the right hand side.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the matrices A_j. On exit, the QR (or LQ) factorizations of A_j as returned by "GEQRF_STRIDED_BATCHED" (or "GELQF_STRIDED_BATCHED").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_j.
[in]strideAhipblasStride.
Stride from the start of one matrix A_j to the next one A_(j+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n
[in,out]Bpointer to type. Array on the GPU (the size depends on the value of strideB).
On entry, the matrices B_j. On exit, when info[j] = 0, each B_j is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrices B_j.
[in]strideBhipblasStride.
Stride from the start of one matrix B_j to the next one B_(j+1). There is no restriction for the value of strideB. Normal use case is strideB >= ldb*nrhs
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int. Array of batchCount integers on the GPU.
If deviceInfo[j] = 0, successful exit for solution of A_j. If deviceInfo[j] = i > 0, the solution of A_j could not be computed because input matrix A_j is rank deficient; the i-th diagonal element of its triangular factor is zero.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasDgemm()

hipblasStatus_t hipblasDgemm ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const double *  alpha,
const double *  AP,
int  lda,
const double *  BP,
int  ldb,
const double *  beta,
double *  CP,
int  ldc 
)

BLAS Level 3 API.

gemm performs one of the matrix-matrix operations

C = alpha*op( A )*op( B ) + beta*C,

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are matrices, with op( A ) an m by k matrix, op( B ) a k by n matrix and C an m by n matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t]
      .
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] number or rows of matrices op( A ) and C
[in]n[int] number of columns of matrices op( B ) and C
[in]k[int] number of columns of matrix op( A ) and number of rows of matrix op( B )
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]BPdevice pointer storing matrix B.
[in]ldb[int] specifies the leading dimension of B.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice pointer storing matrix C on the GPU.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasDgemmBatched()

hipblasStatus_t hipblasDgemmBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const double *  alpha,
const double *const  AP[],
int  lda,
const double *const  BP[],
int  ldb,
const double *  beta,
double *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

gemmBatched performs one of the batched matrix-matrix operations C_i = alpha*op( A_i )*op( B_i ) + beta*C_i, for i = 1, ..., batchCount. where op( X ) is one of op( X ) = X or op( X ) = X**T or op( X ) = X**H, alpha and beta are scalars, and A, B and C are strided batched matrices, with op( A ) an m by k by batchCount strided_batched matrix, op( B ) an k by n by batchCount strided_batched matrix and C an m by n by batchCount strided_batched matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimention m.
[in]n[int] matrix dimention n.
[in]k[int] matrix dimention k.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]BPdevice array of device pointers storing each matrix B_i.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice array of device pointers storing each matrix C_i.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]batchCount[int] number of gemm operations in the batch

◆ hipblasDgemmStridedBatched()

hipblasStatus_t hipblasDgemmStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const double *  alpha,
const double *  AP,
int  lda,
long long  strideA,
const double *  BP,
int  ldb,
long long  strideB,
const double *  beta,
double *  CP,
int  ldc,
long long  strideC,
int  batchCount 
)

BLAS Level 3 API.

gemmStridedBatched performs one of the strided batched matrix-matrix operations

C_i = alpha*op( A_i )*op( B_i ) + beta*C_i, for i = 1, ..., batchCount.

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are strided batched matrices, with op( A ) an m by k by batchCount strided_batched matrix, op( B ) an k by n by batchCount strided_batched matrix and C an m by n by batchCount strided_batched matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimention m.
[in]n[int] matrix dimention n.
[in]k[int] matrix dimention k.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer pointing to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in]BPdevice pointer pointing to the first matrix B_1.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]strideB[hipblasStride] stride from the start of one B_i matrix to the next B_(i + 1).
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice pointer pointing to the first matrix C_1.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]strideC[hipblasStride] stride from the start of one C_i matrix to the next C_(i + 1).
[in]batchCount[int] number of gemm operatons in the batch

◆ hipblasDgemv()

hipblasStatus_t hipblasDgemv ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
const double *  x,
int  incx,
const double *  beta,
double *  y,
int  incy 
)

BLAS Level 2 API.

gemv performs one of the matrix-vector operations

y := alpha*A*x    + beta*y,   or
y := alpha*A**T*x + beta*y,   or
y := alpha*A**H*x + beta*y,

where alpha and beta are scalars, x and y are vectors and A is an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasDgemvBatched()

hipblasStatus_t hipblasDgemvBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
const double *  alpha,
const double *const  AP[],
int  lda,
const double *const  x[],
int  incx,
const double *  beta,
double *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

gemvBatched performs a batch of matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrices A_i are tranposed (conjugated) or not
[in]m[int] number of rows of each matrix A_i
[in]n[int] number of columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each matrix A_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasDgemvStridedBatched()

hipblasStatus_t hipblasDgemvStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
const double *  x,
int  incx,
hipblasStride  stridex,
const double *  beta,
double *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

gemvStridedBatched performs a batch of matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] indicates whether matrices A_i are tranposed (conjugated) or not
[in]m[int] number of rows of matrices A_i
[in]n[int] number of columns of matrices A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer to the first matrix (A_1) in the batch.
[in]lda[int] specifies the leading dimension of matrices A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. When trans equals HIPBLAS_OP_N this typically means stridex >= n * incx, otherwise stridex >= m * incx.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of vectors y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. When trans equals HIPBLAS_OP_N this typically means stridey >= m * incy, otherwise stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasDgeqrf()

hipblasStatus_t hipblasDgeqrf ( hipblasHandle_t  handle,
const int  m,
const int  n,
double *  A,
const int  lda,
double *  ipiv,
int *  info 
)

SOLVER API.

geqrf computes a QR factorization of a general m-by-n matrix A.

The factorization has the form

\[ A = Q\left[\begin{array}{c} R\\ 0 \end{array}\right] \]

where R is upper triangular (upper trapezoidal if m < n), and Q is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q = H_1H_2\cdots H_k, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_i\) is given by

\[ H_i = I - \text{ipiv}[i] \cdot v_i v_i' \]

where the first i-1 elements of the Householder vector \(v_i\) are zero, and \(v_i[i] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of the matrix A.
[in]nint. n >= 0.
The number of columns of the matrix A.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the m-by-n matrix to be factored. On exit, the elements on and above the diagonal contain the factor R; the elements below the diagonal are the last m - i elements of Householder vector v_i.
[in]ldaint. lda >= m.
Specifies the leading dimension of A.
[out]ipivpointer to type. Array on the GPU of dimension min(m,n).
The Householder scalars.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.

◆ hipblasDgeqrfBatched()

hipblasStatus_t hipblasDgeqrfBatched ( hipblasHandle_t  handle,
const int  m,
const int  n,
double *const  A[],
const int  lda,
double *const  ipiv[],
int *  info,
const int  batchCount 
)

SOLVER API.

geqrfBatched computes the QR factorization of a batch of general m-by-n matrices.

The factorization of matrix \(A_i\) in the batch has the form

\[ A_i = Q_i\left[\begin{array}{c} R_i\\ 0 \end{array}\right] \]

where \(R_i\) is upper triangular (upper trapezoidal if m < n), and \(Q_i\) is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q_i = H_{i_1}H_{i_2}\cdots H_{i_k}, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_{i_j}\) is given by

\[ H_{i_j} = I - \text{ipiv}_i[j] \cdot v_{i_j} v_{i_j}' \]

where the first j-1 elements of Householder vector \(v_{i_j}\) are zero, and \(v_{i_j}[j] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of all the matrices A_i in the batch.
[in]nint. n >= 0.
The number of columns of all the matrices A_i in the batch.
[in,out]AArray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the m-by-n matrices A_i to be factored. On exit, the elements on and above the diagonal contain the factor R_i. The elements below the diagonal are the last m - j elements of Householder vector v_(i_j).
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_i.
[out]ipivarray of pointers to type. Each pointer points to an array on the GPU of dimension min(m, n).
Contains the vectors ipiv_i of corresponding Householder scalars.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasDgeqrfStridedBatched()

hipblasStatus_t hipblasDgeqrfStridedBatched ( hipblasHandle_t  handle,
const int  m,
const int  n,
double *  A,
const int  lda,
const hipblasStride  strideA,
double *  ipiv,
const hipblasStride  strideP,
int *  info,
const int  batchCount 
)

SOLVER API.

geqrfStridedBatched computes the QR factorization of a batch of general m-by-n matrices.

The factorization of matrix \(A_i\) in the batch has the form

\[ A_i = Q_i\left[\begin{array}{c} R_i\\ 0 \end{array}\right] \]

where \(R_i\) is upper triangular (upper trapezoidal if m < n), and \(Q_i\) is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q_i = H_{i_1}H_{i_2}\cdots H_{i_k}, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_{i_j}\) is given by

\[ H_{i_j} = I - \text{ipiv}_j[j] \cdot v_{i_j} v_{i_j}' \]

where the first j-1 elements of Householder vector \(v_{i_j}\) are zero, and \(v_{i_j}[j] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of all the matrices A_i in the batch.
[in]nint. n >= 0.
The number of columns of all the matrices A_i in the batch.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the m-by-n matrices A_i to be factored. On exit, the elements on and above the diagonal contain the factor R_i. The elements below the diagonal are the last m - j elements of Householder vector v_(i_j).
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n.
[out]ipivpointer to type. Array on the GPU (the size depends on the value of strideP).
Contains the vectors ipiv_i of corresponding Householder scalars.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use is strideP >= min(m,n).
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasDger()

hipblasStatus_t hipblasDger ( hipblasHandle_t  handle,
int  m,
int  n,
const double *  alpha,
const double *  x,
int  incx,
const double *  y,
int  incy,
double *  AP,
int  lda 
)

BLAS Level 2 API.

ger,geru,gerc performs the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where alpha is a scalar, x and y are vectors, and A is an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of the matrix A.
[in]n[int] the number of columns of the matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasDgerBatched()

hipblasStatus_t hipblasDgerBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const double *  alpha,
const double *const  x[],
int  incx,
const double *const  y[],
int  incy,
double *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

gerBatched,geruBatched,gercBatched performs a batch of the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of eaceh matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasDgerStridedBatched()

hipblasStatus_t hipblasDgerStridedBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const double *  alpha,
const double *  x,
int  incx,
hipblasStride  stridex,
const double *  y,
int  incy,
hipblasStride  stridey,
double *  AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

gerStridedBatched,geruStridedBatched,gercStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**T, OR
A_i := A_i + alpha*x_i*y_i**H  for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increments for the elements of each vector x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size, for a typical case this means stridex >= m * incx.
[in,out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size, for a typical case this means stridey >= n * incy.
[in,out]APdevice pointer to the first matrix (A_1) in the batch.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasDgetrf()

hipblasStatus_t hipblasDgetrf ( hipblasHandle_t  handle,
const int  n,
double *  A,
const int  lda,
int *  ipiv,
int *  info 
)

SOLVER API.

getrf computes the LU factorization of a general n-by-n matrix A using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization has the form:

\[ A = PLU \]

where P is a permutation matrix, L is lower triangular with unit diagonal elements, and U is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A = LU \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of the matrix A.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the n-by-n matrix A to be factored. On exit, the factors L and U from the factorization. The unit diagonal elements of L are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of A.
[out]ipivpointer to int. Array on the GPU of dimension n.
The vector of pivot indices. Elements of ipiv are 1-based indices. For 1 <= i <= n, the row i of the matrix was interchanged with row ipiv[i]. Matrix P of the factorization can be derived from ipiv. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[out]infopointer to a int on the GPU.
If info = 0, successful exit. If info = j > 0, U is singular. U[j,j] is the first zero pivot.

◆ hipblasDgetrfBatched()

hipblasStatus_t hipblasDgetrfBatched ( hipblasHandle_t  handle,
const int  n,
double *const  A[],
const int  lda,
int *  ipiv,
int *  info,
const int  batchCount 
)

SOLVER API.

getrfBatched computes the LU factorization of a batch of general n-by-n matrices using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization of matrix \(A_i\) in the batch has the form:

\[ A_i = P_iL_iU_i \]

where \(P_i\) is a permutation matrix, \(L_i\) is lower triangular with unit diagonal elements, and \(U_i\) is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A_i = L_iU_i \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of all matrices A_i in the batch.
[in,out]Aarray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the n-by-n matrices A_i to be factored. On exit, the factors L_i and U_i from the factorizations. The unit diagonal elements of L_i are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[out]ipivpointer to int. Array on the GPU.
Contains the vectors of pivot indices ipiv_i (corresponding to A_i). Dimension of ipiv_i is n. Elements of ipiv_i are 1-based indices. For each instance A_i in the batch and for 1 <= j <= n, the row j of the matrix A_i was interchanged with row ipiv_i[j]. Matrix P_i of the factorization can be derived from ipiv_i. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for factorization of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasDgetrfStridedBatched()

hipblasStatus_t hipblasDgetrfStridedBatched ( hipblasHandle_t  handle,
const int  n,
double *  A,
const int  lda,
const hipblasStride  strideA,
int *  ipiv,
const hipblasStride  strideP,
int *  info,
const int  batchCount 
)

SOLVER API.

getrfStridedBatched computes the LU factorization of a batch of general n-by-n matrices using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization of matrix \(A_i\) in the batch has the form:

\[ A_i = P_iL_iU_i \]

where \(P_i\) is a permutation matrix, \(L_i\) is lower triangular with unit diagonal elements, and \(U_i\) is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A_i = L_iU_i \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of all matrices A_i in the batch.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the n-by-n matrices A_i to be factored. On exit, the factors L_i and U_i from the factorization. The unit diagonal elements of L_i are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n
[out]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
Contains the vectors of pivots indices ipiv_i (corresponding to A_i). Dimension of ipiv_i is n. Elements of ipiv_i are 1-based indices. For each instance A_i in the batch and for 1 <= j <= n, the row j of the matrix A_i was interchanged with row ipiv_i[j]. Matrix P_i of the factorization can be derived from ipiv_i. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use case is strideP >= n.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for factorization of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasDgetriBatched()

hipblasStatus_t hipblasDgetriBatched ( hipblasHandle_t  handle,
const int  n,
double *const  A[],
const int  lda,
int *  ipiv,
double *const  C[],
const int  ldc,
int *  info,
const int  batchCount 
)

SOLVER API.

getriBatched computes the inverse \(C_i = A_i^{-1}\) of a batch of general n-by-n matrices \(A_i\).

The inverse is computed by solving the linear system

\[ A_i C_i = I \]

where I is the identity matrix, and \(A_i\) is factorized as \(A_i = P_i L_i U_i\) as given by getrfBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of rows and columns of all matrices A_i in the batch.
[in]Aarray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfBatched.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[in]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
The pivot indices returned by getrfBatched. ipiv can be passed in as a nullptr, this will assume that getrfBatched was called without partial pivoting.
[out]Carray of pointers to type. Each pointer points to an array on the GPU of dimension ldc*n.
If info[i] = 0, the inverse of matrices A_i. Otherwise, undefined.
[in]ldcint. ldc >= n.
Specifies the leading dimension of C_i.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for inversion of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasDgetrs()

hipblasStatus_t hipblasDgetrs ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
double *  A,
const int  lda,
const int *  ipiv,
double *  B,
const int  ldb,
int *  info 
)

SOLVER API.

getrs solves a system of n linear equations on n variables in its factorized form.

It solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A X = B & \: \text{not transposed,}\\ A^T X = B & \: \text{transposed, or}\\ A^H X = B & \: \text{conjugate transposed.} \end{array} \]

Matrix A is defined by its triangular factors as returned by getrf.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of A.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of the matrix B.
[in]Apointer to type. Array on the GPU of dimension lda*n.
The factors L and U of the factorization A = P*L*U returned by getrf.
[in]ldaint. lda >= n.
The leading dimension of A.
[in]ipivpointer to int. Array on the GPU of dimension n.
The pivot indices returned by getrf.
[in,out]Bpointer to type. Array on the GPU of dimension ldb*nrhs.
On entry, the right hand side matrix B. On exit, the solution matrix X.
[in]ldbint. ldb >= n.
The leading dimension of B.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.

◆ hipblasDgetrsBatched()

hipblasStatus_t hipblasDgetrsBatched ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
double *const  A[],
const int  lda,
const int *  ipiv,
double *const  B[],
const int  ldb,
int *  info,
const int  batchCount 
)

SOLVER API.

getrsBatched solves a batch of systems of n linear equations on n variables in its factorized forms.

For each instance i in the batch, it solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A_i X_i = B_i & \: \text{not transposed,}\\ A_i^T X_i = B_i & \: \text{transposed, or}\\ A_i^H X_i = B_i & \: \text{conjugate transposed.} \end{array} \]

Matrix \(A_i\) is defined by its triangular factors as returned by getrfBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations of each instance in the batch.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of all A_i matrices.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of all the matrices B_i.
[in]AArray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfBatched.
[in]ldaint. lda >= n.
The leading dimension of matrices A_i.
[in]ipivpointer to int. Array on the GPU.
Contains the vectors ipiv_i of pivot indices returned by getrfBatched.
[in,out]BArray of pointers to type. Each pointer points to an array on the GPU of dimension ldb*nrhs.
On entry, the right hand side matrices B_i. On exit, the solution matrix X_i of each system in the batch.
[in]ldbint. ldb >= n.
The leading dimension of matrices B_i.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of instances (systems) in the batch.

◆ hipblasDgetrsStridedBatched()

hipblasStatus_t hipblasDgetrsStridedBatched ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
double *  A,
const int  lda,
const hipblasStride  strideA,
const int *  ipiv,
const hipblasStride  strideP,
double *  B,
const int  ldb,
const hipblasStride  strideB,
int *  info,
const int  batchCount 
)

SOLVER API.

getrsStridedBatched solves a batch of systems of n linear equations on n variables in its factorized forms.

For each instance i in the batch, it solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A_i X_i = B_i & \: \text{not transposed,}\\ A_i^T X_i = B_i & \: \text{transposed, or}\\ A_i^H X_i = B_i & \: \text{conjugate transposed.} \end{array} \]

Matrix \(A_i\) is defined by its triangular factors as returned by getrfStridedBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations of each instance in the batch.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of all A_i matrices.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of all the matrices B_i.
[in]Apointer to type. Array on the GPU (the size depends on the value of strideA).
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfStridedBatched.
[in]ldaint. lda >= n.
The leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n.
[in]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
Contains the vectors ipiv_i of pivot indices returned by getrfStridedBatched.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use case is strideP >= n.
[in,out]Bpointer to type. Array on the GPU (size depends on the value of strideB).
On entry, the right hand side matrices B_i. On exit, the solution matrix X_i of each system in the batch.
[in]ldbint. ldb >= n.
The leading dimension of matrices B_i.
[in]strideBhipblasStride.
Stride from the start of one matrix B_i to the next one B_(i+1). There is no restriction for the value of strideB. Normal use case is strideB >= ldb*nrhs.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of instances (systems) in the batch.

◆ hipblasDnrm2()

hipblasStatus_t hipblasDnrm2 ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
double *  result 
)

BLAS Level 1 API.

nrm2 computes the euclidean norm of a real or complex vector

      result := sqrt( x'*x ) for real vectors
      result := sqrt( x**H*x ) for complex vectors
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : s,d,sc,dz
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the nrm2 product. return is 0.0 if n, incx<=0.

◆ hipblasDnrm2Batched()

hipblasStatus_t hipblasDnrm2Batched ( hipblasHandle_t  handle,
int  n,
const double *const  x[],
int  incx,
int  batchCount,
double *  result 
)

BLAS Level 1 API.

nrm2Batched computes the euclidean norm over a batch of real or complex vectors

      result := sqrt( x_i'*x_i ) for real vectors x, for i = 1, ..., batchCount
      result := sqrt( x_i**H*x_i ) for complex vectors x, for i = 1, ..., batchCount
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array of batchCount size for nrm2 results. return is 0.0 for each element if n <= 0, incx<=0.

◆ hipblasDnrm2StridedBatched()

hipblasStatus_t hipblasDnrm2StridedBatched ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
hipblasStride  stridex,
int  batchCount,
double *  result 
)

BLAS Level 1 API.

nrm2StridedBatched computes the euclidean norm over a batch of real or complex vectors

      := sqrt( x_i'*x_i ) for real vectors x, for i = 1, ..., batchCount
      := sqrt( x_i**H*x_i ) for complex vectors, for i = 1, ..., batchCount
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array for storing contiguous batchCount results. return is 0.0 for each element if n <= 0, incx<=0.

◆ hipblasDotBatchedEx()

hipblasStatus_t hipblasDotBatchedEx ( hipblasHandle_t  handle,
int  n,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
const void *  y,
hipblasDatatype_t  yType,
int  incy,
int  batchCount,
void *  result,
hipblasDatatype_t  resultType,
hipblasDatatype_t  executionType 
)

BLAS EX API.

dotBatchedEx performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcBatchedEx performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

- Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]yType[hipblasDatatype_t] specifies the datatype of each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.
[in]resultType[hipblasDatatype_t] specifies the datatype of the result.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasDotcBatchedEx()

hipblasStatus_t hipblasDotcBatchedEx ( hipblasHandle_t  handle,
int  n,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
const void *  y,
hipblasDatatype_t  yType,
int  incy,
int  batchCount,
void *  result,
hipblasDatatype_t  resultType,
hipblasDatatype_t  executionType 
)

BLAS EX API.

dotBatchedEx performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcBatchedEx performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

- Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]yType[hipblasDatatype_t] specifies the datatype of each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.
[in]resultType[hipblasDatatype_t] specifies the datatype of the result.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasDotcEx()

hipblasStatus_t hipblasDotcEx ( hipblasHandle_t  handle,
int  n,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
const void *  y,
hipblasDatatype_t  yType,
int  incy,
void *  result,
hipblasDatatype_t  resultType,
hipblasDatatype_t  executionType 
)

BLAS EX API.

dotEx performs the dot product of vectors x and y

result = x * y;

dotcEx performs the dot product of the conjugate of complex vector x and complex vector y

result = conjugate (x) * y;

- Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]xType[hipblasDatatype_t] specifies the datatype of vector x.
[in]incx[int] specifies the increment for the elements of y.
[in]ydevice pointer storing vector y.
[in]yType[hipblasDatatype_t] specifies the datatype of vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the dot product. return is 0.0 if n <= 0.
[in]resultType[hipblasDatatype_t] specifies the datatype of the result.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasDotcStridedBatchedEx()

hipblasStatus_t hipblasDotcStridedBatchedEx ( hipblasHandle_t  handle,
int  n,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
hipblasStride  stridex,
const void *  y,
hipblasDatatype_t  yType,
int  incy,
hipblasStride  stridey,
int  batchCount,
void *  result,
hipblasDatatype_t  resultType,
hipblasDatatype_t  executionType 
)

BLAS EX API.

dotStridedBatchedEx performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotc_strided_batched_ex performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

- Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]ydevice pointer to the first vector (y_1) in the batch.
[in]yType[hipblasDatatype_t] specifies the datatype of each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.
[in]resultType[hipblasDatatype_t] specifies the datatype of the result.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasDotEx()

hipblasStatus_t hipblasDotEx ( hipblasHandle_t  handle,
int  n,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
const void *  y,
hipblasDatatype_t  yType,
int  incy,
void *  result,
hipblasDatatype_t  resultType,
hipblasDatatype_t  executionType 
)

BLAS EX API.

dotEx performs the dot product of vectors x and y

result = x * y;

dotcEx performs the dot product of the conjugate of complex vector x and complex vector y

result = conjugate (x) * y;

- Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]xType[hipblasDatatype_t] specifies the datatype of vector x.
[in]incx[int] specifies the increment for the elements of y.
[in]ydevice pointer storing vector y.
[in]yType[hipblasDatatype_t] specifies the datatype of vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the dot product. return is 0.0 if n <= 0.
[in]resultType[hipblasDatatype_t] specifies the datatype of the result.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasDotStridedBatchedEx()

hipblasStatus_t hipblasDotStridedBatchedEx ( hipblasHandle_t  handle,
int  n,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
hipblasStride  stridex,
const void *  y,
hipblasDatatype_t  yType,
int  incy,
hipblasStride  stridey,
int  batchCount,
void *  result,
hipblasDatatype_t  resultType,
hipblasDatatype_t  executionType 
)

BLAS EX API.

dotStridedBatchedEx performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotc_strided_batched_ex performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

- Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]ydevice pointer to the first vector (y_1) in the batch.
[in]yType[hipblasDatatype_t] specifies the datatype of each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.
[in]resultType[hipblasDatatype_t] specifies the datatype of the result.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasDrot()

hipblasStatus_t hipblasDrot ( hipblasHandle_t  handle,
int  n,
double *  x,
int  incx,
double *  y,
int  incy,
const double *  c,
const double *  s 
)

BLAS Level 1 API.

rot applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to vectors x and y. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment between elements of y.
[in]cdevice pointer or host pointer storing scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer storing scalar sine component of the rotation matrix.

◆ hipblasDrotBatched()

hipblasStatus_t hipblasDrotBatched ( hipblasHandle_t  handle,
int  n,
double *const  x[],
int  incx,
double *const  y[],
int  incy,
const double *  c,
const double *  s,
int  batchCount 
)

BLAS Level 1 API.

rotBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice array of deivce pointers storing each vector x_i.
[in]incx[int] specifies the increment between elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment between elements of each y_i.
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasDrotg()

hipblasStatus_t hipblasDrotg ( hipblasHandle_t  handle,
double *  a,
double *  b,
double *  c,
double *  s 
)

BLAS Level 1 API.

rotg creates the Givens rotation matrix for the vector (a b). Scalars c and s and arrays a and b may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice pointer or host pointer to input vector element, overwritten with r.
[in,out]bdevice pointer or host pointer to input vector element, overwritten with z.
[in,out]cdevice pointer or host pointer to cosine element of Givens rotation.
[in,out]sdevice pointer or host pointer sine element of Givens rotation.

◆ hipblasDrotgBatched()

hipblasStatus_t hipblasDrotgBatched ( hipblasHandle_t  handle,
double *const  a[],
double *const  b[],
double *const  c[],
double *const  s[],
int  batchCount 
)

BLAS Level 1 API.

rotgBatched creates the Givens rotation matrix for the batched vectors (a_i b_i), for i = 1, ..., batchCount. a, b, c, and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice array of device pointers storing each single input vector element a_i, overwritten with r_i.
[in,out]bdevice array of device pointers storing each single input vector element b_i, overwritten with z_i.
[in,out]cdevice array of device pointers storing each cosine element of Givens rotation for the batch.
[in,out]sdevice array of device pointers storing each sine element of Givens rotation for the batch.
[in]batchCount[int] number of batches (length of arrays a, b, c, and s).

◆ hipblasDrotgStridedBatched()

hipblasStatus_t hipblasDrotgStridedBatched ( hipblasHandle_t  handle,
double *  a,
hipblasStride  stridea,
double *  b,
hipblasStride  strideb,
double *  c,
hipblasStride  stridec,
double *  s,
hipblasStride  strides,
int  batchCount 
)

BLAS Level 1 API.

rotgStridedBatched creates the Givens rotation matrix for the strided batched vectors (a_i b_i), for i = 1, ..., batchCount. a, b, c, and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice strided_batched pointer or host strided_batched pointer to first single input vector element a_1, overwritten with r.
[in]stridea[hipblasStride] distance between elements of a in batch (distance between a_i and a_(i + 1))
[in,out]bdevice strided_batched pointer or host strided_batched pointer to first single input vector element b_1, overwritten with z.
[in]strideb[hipblasStride] distance between elements of b in batch (distance between b_i and b_(i + 1))
[in,out]cdevice strided_batched pointer or host strided_batched pointer to first cosine element of Givens rotations c_1.
[in]stridec[hipblasStride] distance between elements of c in batch (distance between c_i and c_(i + 1))
[in,out]sdevice strided_batched pointer or host strided_batched pointer to sine element of Givens rotations s_1.
[in]strides[hipblasStride] distance between elements of s in batch (distance between s_i and s_(i + 1))
[in]batchCount[int] number of batches (length of arrays a, b, c, and s).

◆ hipblasDrotm()

hipblasStatus_t hipblasDrotm ( hipblasHandle_t  handle,
int  n,
double *  x,
int  incx,
double *  y,
int  incy,
const double *  param 
)

BLAS Level 1 API.

rotm applies the modified Givens rotation matrix defined by param to vectors x and y.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : s,d
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment between elements of y.
[in]paramdevice vector or host vector of 5 elements defining the rotation. param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

◆ hipblasDrotmBatched()

hipblasStatus_t hipblasDrotmBatched ( hipblasHandle_t  handle,
int  n,
double *const  x[],
int  incx,
double *const  y[],
int  incy,
const double *const  param[],
int  batchCount 
)

BLAS Level 1 API.

rotmBatched applies the modified Givens rotation matrix defined by param_i to batched vectors x_i and y_i, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment between elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_1.
[in]incy[int] specifies the increment between elements of each y_i.
[in]paramdevice array of device vectors of 5 elements defining the rotation. param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may ONLY be stored on the device for the batched version of this function.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasDrotmg()

hipblasStatus_t hipblasDrotmg ( hipblasHandle_t  handle,
double *  d1,
double *  d2,
double *  x1,
const double *  y1,
double *  param 
)

BLAS Level 1 API.

rotmg creates the modified Givens rotation matrix for the vector (d1 * x1, d2 * y1). Parameters may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : s,d
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]d1device pointer or host pointer to input scalar that is overwritten.
[in,out]d2device pointer or host pointer to input scalar that is overwritten.
[in,out]x1device pointer or host pointer to input scalar that is overwritten.
[in]y1device pointer or host pointer to input scalar.
[out]paramdevice vector or host vector of 5 elements defining the rotation. param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

◆ hipblasDrotmgBatched()

hipblasStatus_t hipblasDrotmgBatched ( hipblasHandle_t  handle,
double *const  d1[],
double *const  d2[],
double *const  x1[],
const double *const  y1[],
double *const  param[],
int  batchCount 
)

BLAS Level 1 API.

rotmgBatched creates the modified Givens rotation matrix for the batched vectors (d1_i * x1_i, d2_i * y1_i), for i = 1, ..., batchCount. Parameters may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]d1device batched array or host batched array of input scalars that is overwritten.
[in,out]d2device batched array or host batched array of input scalars that is overwritten.
[in,out]x1device batched array or host batched array of input scalars that is overwritten.
[in]y1device batched array or host batched array of input scalars.
[out]paramdevice batched array or host batched array of vectors of 5 elements defining the rotation. param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.
[in]batchCount[int] the number of instances in the batch.

◆ hipblasDrotmgStridedBatched()

hipblasStatus_t hipblasDrotmgStridedBatched ( hipblasHandle_t  handle,
double *  d1,
hipblasStride  strided1,
double *  d2,
hipblasStride  strided2,
double *  x1,
hipblasStride  stridex1,
const double *  y1,
hipblasStride  stridey1,
double *  param,
hipblasStride  strideParam,
int  batchCount 
)

BLAS Level 1 API.

rotmgStridedBatched creates the modified Givens rotation matrix for the strided batched vectors (d1_i * x1_i, d2_i * y1_i), for i = 1, ..., batchCount. Parameters may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]d1device strided_batched array or host strided_batched array of input scalars that is overwritten.
[in]strided1[hipblasStride] specifies the increment between the beginning of d1_i and d1_(i+1)
[in,out]d2device strided_batched array or host strided_batched array of input scalars that is overwritten.
[in]strided2[hipblasStride] specifies the increment between the beginning of d2_i and d2_(i+1)
[in,out]x1device strided_batched array or host strided_batched array of input scalars that is overwritten.
[in]stridex1[hipblasStride] specifies the increment between the beginning of x1_i and x1_(i+1)
[in]y1device strided_batched array or host strided_batched array of input scalars.
[in]stridey1[hipblasStride] specifies the increment between the beginning of y1_i and y1_(i+1)
[out]paramdevice stridedBatched array or host stridedBatched array of vectors of 5 elements defining the rotation. param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.
[in]strideParam[hipblasStride] specifies the increment between the beginning of param_i and param_(i + 1)
[in]batchCount[int] the number of instances in the batch.

◆ hipblasDrotmStridedBatched()

hipblasStatus_t hipblasDrotmStridedBatched ( hipblasHandle_t  handle,
int  n,
double *  x,
int  incx,
hipblasStride  stridex,
double *  y,
int  incy,
hipblasStride  stridey,
const double *  param,
hipblasStride  strideParam,
int  batchCount 
)

BLAS Level 1 API.

rotmStridedBatched applies the modified Givens rotation matrix defined by param_i to strided batched vectors x_i and y_i, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice pointer pointing to first strided batched vector x_1.
[in]incx[int] specifies the increment between elements of each x_i.
[in]stridex[hipblasStride] specifies the increment between the beginning of x_i and x_(i + 1)
[in,out]ydevice pointer pointing to first strided batched vector y_1.
[in]incy[int] specifies the increment between elements of each y_i.
[in]stridey[hipblasStride] specifies the increment between the beginning of y_i and y_(i + 1)
[in]paramdevice pointer pointing to first array of 5 elements defining the rotation (param_1). param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may ONLY be stored on the device for the strided_batched version of this function.
[in]strideParam[hipblasStride] specifies the increment between the beginning of param_i and param_(i + 1)
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasDrotStridedBatched()

hipblasStatus_t hipblasDrotStridedBatched ( hipblasHandle_t  handle,
int  n,
double *  x,
int  incx,
hipblasStride  stridex,
double *  y,
int  incy,
hipblasStride  stridey,
const double *  c,
const double *  s,
int  batchCount 
)

BLAS Level 1 API.

rotStridedBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to strided batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment between elements of each x_i.
[in]stridex[hipblasStride] specifies the increment from the beginning of x_i to the beginning of x_(i+1)
[in,out]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment between elements of each y_i.
[in]stridey[hipblasStride] specifies the increment from the beginning of y_i to the beginning of y_(i+1)
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasDsbmv()

hipblasStatus_t hipblasDsbmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const double *  alpha,
const double *  AP,
int  lda,
const double *  x,
int  incx,
const double *  beta,
double *  y,
int  incy 
)

BLAS Level 2 API.

sbmv performs the matrix-vector operation:

y := alpha*A*x + beta*y,

where alpha and beta are scalars, x and y are n element vectors and A should contain an upper or lower triangular n by n symmetric banded matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : s,d
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int]
[in]k[int] specifies the number of sub- and super-diagonals
[in]alphaspecifies the scalar alpha
[in]APpointer storing matrix A on the GPU
[in]lda[int] specifies the leading dimension of matrix A
[in]xpointer storing vector x on the GPU
[in]incx[int] specifies the increment for the elements of x
[in]betaspecifies the scalar beta
[out]ypointer storing vector y on the GPU
[in]incy[int] specifies the increment for the elements of y

◆ hipblasDsbmvBatched()

hipblasStatus_t hipblasDsbmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const double *  alpha,
const double *const  AP[],
int  lda,
const double *const  x[],
int  incx,
const double *  beta,
double *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

sbmvBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric banded matrix, for i = 1, ..., batchCount. A should contain an upper or lower triangular n by n symmetric banded matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]k[int] specifies the number of sub- and super-diagonals
[in]alphadevice pointer or host pointer to scalar alpha
[in]APdevice array of device pointers storing each matrix A_i
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]xdevice array of device pointers storing each vector x_i
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]betadevice pointer or host pointer to scalar beta
[out]ydevice array of device pointers storing each vector y_i
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]batchCount[int] number of instances in the batch

◆ hipblasDsbmvStridedBatched()

hipblasStatus_t hipblasDsbmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
const double *  x,
int  incx,
hipblasStride  stridex,
const double *  beta,
double *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

sbmvStridedBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric banded matrix, for i = 1, ..., batchCount. A should contain an upper or lower triangular n by n symmetric banded matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]k[int] specifies the number of sub- and super-diagonals
[in]alphadevice pointer or host pointer to scalar alpha
[in]APDevice pointer to the first matrix A_1 on the GPU
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xDevice pointer to the first vector x_1 on the GPU
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. This typically means stridex >= n * incx. stridex should be non zero.
[in]betadevice pointer or host pointer to scalar beta
[out]yDevice pointer to the first vector y_1 on the GPU
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. This typically means stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasDscal()

hipblasStatus_t hipblasDscal ( hipblasHandle_t  handle,
int  n,
const double *  alpha,
double *  x,
int  incx 
)

BLAS Level 1 API.

scal scales each element of vector x with scalar alpha.

x := alpha * x
  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]alphadevice pointer or host pointer for the scalar alpha.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasDscalBatched()

hipblasStatus_t hipblasDscalBatched ( hipblasHandle_t  handle,
int  n,
const double *  alpha,
double *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 1 API.

scalBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasDscalStridedBatched()

hipblasStatus_t hipblasDscalStridedBatched ( hipblasHandle_t  handle,
int  n,
const double *  alpha,
double *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 1 API.

scalStridedBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i ,

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasDspmv()

hipblasStatus_t hipblasDspmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  AP,
const double *  x,
int  incx,
const double *  beta,
double *  y,
int  incy 
)

BLAS Level 2 API.

spmv performs the matrix-vector operation:

y := alpha*A*x + beta*y,

where alpha and beta are scalars, x and y are n element vectors and A should contain an upper or lower triangular n by n packed symmetric matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : s,d
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int]
[in]alphaspecifies the scalar alpha
[in]APpointer storing matrix A on the GPU
[in]xpointer storing vector x on the GPU
[in]incx[int] specifies the increment for the elements of x
[in]betaspecifies the scalar beta
[out]ypointer storing vector y on the GPU
[in]incy[int] specifies the increment for the elements of y

◆ hipblasDspmvBatched()

hipblasStatus_t hipblasDspmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *const  AP[],
const double *const  x[],
int  incx,
const double *  beta,
double *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

spmvBatched performs the matrix-vector operation:

y_i := alpha*AP_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A should contain an upper or lower triangular n by n packed symmetric matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APdevice array of device pointers storing each matrix A_i
[in]xdevice array of device pointers storing each vector x_i
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]betadevice pointer or host pointer to scalar beta
[out]ydevice array of device pointers storing each vector y_i
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]batchCount[int] number of instances in the batch

◆ hipblasDspmvStridedBatched()

hipblasStatus_t hipblasDspmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  AP,
hipblasStride  strideA,
const double *  x,
int  incx,
hipblasStride  stridex,
const double *  beta,
double *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

spmvStridedBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A should contain an upper or lower triangular n by n packed symmetric matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APDevice pointer to the first matrix A_1 on the GPU
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xDevice pointer to the first vector x_1 on the GPU
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. This typically means stridex >= n * incx. stridex should be non zero.
[in]betadevice pointer or host pointer to scalar beta
[out]yDevice pointer to the first vector y_1 on the GPU
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. This typically means stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasDspr()

hipblasStatus_t hipblasDspr ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  x,
int  incx,
double *  AP 
)

BLAS Level 2 API.

spr performs the matrix-vector operations

A := A + alpha*x*x**T

where alpha is a scalar, x is a vector, and A is an n by n symmetric matrix, supplied in packed form.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in AP.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing the packed version of the specified triangular portion of the symmetric matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0

◆ hipblasDspr2()

hipblasStatus_t hipblasDspr2 ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  x,
int  incx,
const double *  y,
int  incy,
double *  AP 
)

BLAS Level 2 API.

spr2 performs the matrix-vector operation

A := A + alpha*x*y**T + alpha*y*x**T

where alpha is a scalar, x and y are vectors, and A is an n by n symmetric matrix, supplied in packed form.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : s,d
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in AP.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing the packed version of the specified triangular portion of the symmetric matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(n) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0

◆ hipblasDspr2Batched()

hipblasStatus_t hipblasDspr2Batched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *const  x[],
int  incx,
const double *const  y[],
int  incy,
double *const  AP[],
int  batchCount 
)

BLAS Level 2 API.

spr2Batched performs the matrix-vector operation

A_i := A_i + alpha*x_i*y_i**T + alpha*y_i*x_i**T

where alpha is a scalar, x_i and y_i are vectors, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each symmetric matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(n) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]batchCount[int] number of instances in the batch.

◆ hipblasDspr2StridedBatched()

hipblasStatus_t hipblasDspr2StridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  x,
int  incx,
hipblasStride  stridex,
const double *  y,
int  incy,
hipblasStride  stridey,
double *  AP,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

spr2StridedBatched performs the matrix-vector operation

A_i := A_i + alpha*x_i*y_i**T + alpha*y_i*x_i**T

where alpha is a scalar, x_i amd y_i are vectors, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in]ydevice pointer pointing to the first vector (y_1).
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1).
[in,out]APdevice pointer storing the packed version of the specified triangular portion of each symmetric matrix A_i. Points to the first A_1. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(n) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasDsprBatched()

hipblasStatus_t hipblasDsprBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *const  x[],
int  incx,
double *const  AP[],
int  batchCount 
)

BLAS Level 2 API.

sprBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**T

where alpha is a scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each symmetric matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]batchCount[int] number of instances in the batch.

◆ hipblasDsprStridedBatched()

hipblasStatus_t hipblasDsprStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  x,
int  incx,
hipblasStride  stridex,
double *  AP,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

sprStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**T

where alpha is a scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in,out]APdevice pointer storing the packed version of the specified triangular portion of each symmetric matrix A_i. Points to the first A_1. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasDswap()

hipblasStatus_t hipblasDswap ( hipblasHandle_t  handle,
int  n,
double *  x,
int  incx,
double *  y,
int  incy 
)

BLAS Level 1 API.

swap interchanges vectors x and y.

y := x; x := y
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasDswapBatched()

hipblasStatus_t hipblasDswapBatched ( hipblasHandle_t  handle,
int  n,
double *const  x[],
int  incx,
double *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

swapBatched interchanges vectors x_i and y_i, for i = 1 , ... , batchCount

y_i := x_i; x_i := y_i
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasDswapStridedBatched()

hipblasStatus_t hipblasDswapStridedBatched ( hipblasHandle_t  handle,
int  n,
double *  x,
int  incx,
hipblasStride  stridex,
double *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

swapStridedBatched interchanges vectors x_i and y_i, for i = 1 , ... , batchCount

y_i := x_i; x_i := y_i
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in,out]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in,out]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_y is of appropriate size, for a typical case this means stride_y >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch.

◆ hipblasDsymm()

hipblasStatus_t hipblasDsymm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
const double *  BP,
int  ldb,
const double *  beta,
double *  CP,
int  ldc 
)

BLAS Level 3 API.

symm performs one of the matrix-matrix operations:

C := alpha*A*B + beta*C if side == HIPBLAS_SIDE_LEFT, C := alpha*B*A + beta*C if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B and C are m by n matrices, and A is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C := alpha*A*B + beta*C HIPBLAS_SIDE_RIGHT: C := alpha*B*A + beta*C
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix
[in]m[int] m specifies the number of rows of B and C. m >= 0.
[in]n[int] n specifies the number of columns of B and C. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A and B are not referenced.
[in]APpointer storing matrix A on the GPU. A is m by m if side == HIPBLAS_SIDE_LEFT A is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPpointer storing matrix B on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m )

◆ hipblasDsymmBatched()

hipblasStatus_t hipblasDsymmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const double *  alpha,
const double *const  AP[],
int  lda,
const double *const  BP[],
int  ldb,
const double *  beta,
double *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

symmBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]m[int] m specifies the number of rows of B_i and C_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i and C_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPdevice array of device pointers storing each matrix B_i on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C_i need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C_i. ldc >= max( 1, m )
[in]batchCount[int] number of instances in the batch.

◆ hipblasDsymmStridedBatched()

hipblasStatus_t hipblasDsymmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
const double *  BP,
int  ldb,
hipblasStride  strideB,
const double *  beta,
double *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

symmStridedBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]m[int] m specifies the number of rows of B_i and C_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i and C_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice pointer to first matrix A_1 A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPdevice pointer to first matrix B_1 of dimension (ldb, n) on the GPU.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m )
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice pointer to first matrix C_1 of dimension (ldc, n) on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasDsymv()

hipblasStatus_t hipblasDsymv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
const double *  x,
int  incx,
const double *  beta,
double *  y,
int  incy 
)

BLAS Level 2 API.

symv performs the matrix-vector operation:

y := alpha*A*x + beta*y,

where alpha and beta are scalars, x and y are n element vectors and A should contain an upper or lower triangular n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int]
[in]alphaspecifies the scalar alpha
[in]APpointer storing matrix A on the GPU
[in]lda[int] specifies the leading dimension of A
[in]xpointer storing vector x on the GPU
[in]incx[int] specifies the increment for the elements of x
[in]betaspecifies the scalar beta
[out]ypointer storing vector y on the GPU
[in]incy[int] specifies the increment for the elements of y

◆ hipblasDsymvBatched()

hipblasStatus_t hipblasDsymvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *const  AP[],
int  lda,
const double *const  x[],
int  incx,
const double *  beta,
double *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

symvBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A a should contain an upper or lower triangular symmetric matrix and the opposing triangular part of A is not referenced

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APdevice array of device pointers storing each matrix A_i
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]xdevice array of device pointers storing each vector x_i
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]betadevice pointer or host pointer to scalar beta
[out]ydevice array of device pointers storing each vector y_i
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]batchCount[int] number of instances in the batch

◆ hipblasDsymvStridedBatched()

hipblasStatus_t hipblasDsymvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
const double *  x,
int  incx,
hipblasStride  stridex,
const double *  beta,
double *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

symvStridedBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A a should contain an upper or lower triangular symmetric matrix and the opposing triangular part of A is not referenced

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APDevice pointer to the first matrix A_1 on the GPU
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xDevice pointer to the first vector x_1 on the GPU
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. This typically means stridex >= n * incx. stridex should be non zero.
[in]betadevice pointer or host pointer to scalar beta
[out]yDevice pointer to the first vector y_1 on the GPU
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. This typically means stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasDsyr()

hipblasStatus_t hipblasDsyr ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  x,
int  incx,
double *  AP,
int  lda 
)

BLAS Level 2 API.

syr performs the matrix-vector operations

A := A + alpha*x*x**T

where alpha is a scalar, x is a vector, and A is an n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasDsyr2()

hipblasStatus_t hipblasDsyr2 ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  x,
int  incx,
const double *  y,
int  incy,
double *  AP,
int  lda 
)

BLAS Level 2 API.

syr2 performs the matrix-vector operations

A := A + alpha*x*y**T + alpha*y*x**T

where alpha is a scalar, x and y are vectors, and A is an n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasDsyr2Batched()

hipblasStatus_t hipblasDsyr2Batched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *const  x[],
int  incx,
const double *const  y[],
int  incy,
double *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

syr2Batched performs a batch of matrix-vector operations

A[i] := A[i] + alpha*x[i]*y[i]**T + alpha*y[i]*x[i]**T

where alpha is a scalar, x[i] and y[i] are vectors, and A[i] is a n by n symmetric matrix, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasDsyr2k()

hipblasStatus_t hipblasDsyr2k ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const double *  AP,
int  lda,
const double *  BP,
int  ldb,
const double *  beta,
double *  CP,
int  ldc 
)

BLAS Level 3 API.

syr2k performs one of the matrix-matrix operations for a symmetric rank-2k update

C := alpha*(op( A )*op( B )^T + op( B )*op( A )^T) + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^T, op( B ) = B^T,  and A and B are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A ) = A^T, op( B ) = B^T HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A) and op(B). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasDsyr2kBatched()

hipblasStatus_t hipblasDsyr2kBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const double *const  AP[],
int  lda,
const double *const  BP[],
int  ldb,
const double *  beta,
double *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syr2kBatched performs a batch of the matrix-matrix operations for a symmetric rank-2k update

C_i := alpha*(op( A_i )*op( B_i )^T + op( B_i )*op( A_i )^T) + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasDsyr2kStridedBatched()

hipblasStatus_t hipblasDsyr2kStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
const double *  BP,
int  ldb,
hipblasStride  strideB,
const double *  beta,
double *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syr2kStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-2k update

C_i := alpha*(op( A_i )*op( B_i )^T + op( B_i )*op( A_i )^T) + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasDsyr2StridedBatched()

hipblasStatus_t hipblasDsyr2StridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  x,
int  incx,
hipblasStride  stridex,
const double *  y,
int  incy,
hipblasStride  stridey,
double *  AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

syr2StridedBatched the matrix-vector operations

A[i] := A[i] + alpha*x[i]*y[i]**T + alpha*y[i]*x[i]**T

where alpha is a scalar, x[i] and y[i] are vectors, and A[i] is a n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of each matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the pointer increment between vectors (x_i) and (x_i+1).
[in]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] specifies the pointer increment between vectors (y_i) and (y_i+1).
[in,out]APdevice pointer to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasDsyrBatched()

hipblasStatus_t hipblasDsyrBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *const  x[],
int  incx,
double *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

syrBatched performs a batch of matrix-vector operations

A[i] := A[i] + alpha*x[i]*x[i]**T

where alpha is a scalar, x is an array of vectors, and A is an array of n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasDsyrk()

hipblasStatus_t hipblasDsyrk ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const double *  AP,
int  lda,
const double *  beta,
double *  CP,
int  ldc 
)

BLAS Level 3 API.

syrk performs one of the matrix-matrix operations for a symmetric rank-k update

C := alpha*op( A )*op( A )^T + beta*C

where alpha and beta are scalars, op(A) is an n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower.

op( A ) = A, and A is n by k if transA == HIPBLAS_OP_N
op( A ) = A^T and A is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if transA = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasDsyrkBatched()

hipblasStatus_t hipblasDsyrkBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const double *const  AP[],
int  lda,
const double *  beta,
double *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syrkBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( A_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) is an n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^T and A_i is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasDsyrkStridedBatched()

hipblasStatus_t hipblasDsyrkStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
const double *  beta,
double *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syrkStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( A_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) is an n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^T and A_i is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU. on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasDsyrkx()

hipblasStatus_t hipblasDsyrkx ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const double *  AP,
int  lda,
const double *  BP,
int  ldb,
const double *  beta,
double *  CP,
int  ldc 
)

BLAS Level 3 API.

syrkx performs one of the matrix-matrix operations for a symmetric rank-k update

C := alpha*op( A )*op( B )^T + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A )*op( B )^T will be symmetric.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^T, op( B ) = B^T,  and A and B are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A ) = A^T, op( B ) = B^T HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A) and op(B). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasDsyrkxBatched()

hipblasStatus_t hipblasDsyrkxBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const double *const  AP[],
int  lda,
const double *const  BP[],
int  ldb,
const double *  beta,
double *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syrkxBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( B_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A_i )*op( B_i )^T will be symmetric.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasDsyrkxStridedBatched()

hipblasStatus_t hipblasDsyrkxStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
const double *  BP,
int  ldb,
hipblasStride  strideB,
const double *  beta,
double *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syrkxStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( B_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A_i )*op( B_i )^T will be symmetric.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasDsyrStridedBatched()

hipblasStatus_t hipblasDsyrStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const double *  x,
int  incx,
hipblasStride  stridex,
double *  AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

syrStridedBatched performs the matrix-vector operations

A[i] := A[i] + alpha*x[i]*x[i]**T

where alpha is a scalar, vectors, and A is an array of n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of each matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the pointer increment between vectors (x_i) and (x_i+1).
[in,out]APdevice pointer to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasDtbmv()

hipblasStatus_t hipblasDtbmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const double *  AP,
int  lda,
double *  x,
int  incx 
)

BLAS Level 2 API.

tbmv performs one of the matrix-vector operations

x := A*x      or
x := A**T*x   or
x := A**H*x,

x is a vectors and A is a banded m by m matrix (see description below).

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of A is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of A's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by A.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of the matrix A. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of the matrix A. k must satisfy k > 0 && k < lda.
[in]APdevice pointer storing banded triangular matrix A. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of A. lda must satisfy lda > k.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasDtbmvBatched()

hipblasStatus_t hipblasDtbmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const double *const  AP[],
int  lda,
double *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tbmvBatched performs one of the matrix-vector operations

x_i := A_i*x_i      or
x_i := A_i**T*x_i   or
x_i := A_i**H*x_i,

where (A_i, x_i) is the i-th instance of the batch. x_i is a vector and A_i is an m by m matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether each matrix A_i is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of each A_i is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of each A_i's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by each A_i.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of each matrix A_i. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of each matrix A_i. k must satisfy k > 0 && k < lda.
[in]APdevice array of device pointers storing each banded triangular matrix A_i. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of each A_i. lda must satisfy lda > k.
[in,out]xdevice array of device pointer storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasDtbmvStridedBatched()

hipblasStatus_t hipblasDtbmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const double *  AP,
int  lda,
hipblasStride  strideA,
double *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tbmvStridedBatched performs one of the matrix-vector operations

x_i := A_i*x_i      or
x_i := A_i**T*x_i   or
x_i := A_i**H*x_i,

where (A_i, x_i) is the i-th instance of the batch. x_i is a vector and A_i is an m by m matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether each matrix A_i is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of each A_i is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of each A_i's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by each A_i.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of each matrix A_i. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of each matrix A_i. k must satisfy k > 0 && k < lda.
[in]APdevice array to the first matrix A_i of the batch. Stores each banded triangular matrix A_i. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of each A_i. lda must satisfy lda > k.
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in,out]xdevice array to the first vector x_i of the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one x_i matrix to the next x_(i + 1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasDtbsv()

hipblasStatus_t hipblasDtbsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const double *  AP,
int  lda,
double *  x,
int  incx 
)

BLAS Level 2 API.

tbsv solves

 A*x = b or A**T*x = b or A**H*x = b,

where x and b are vectors and A is a banded triangular matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular (i.e. the diagonal elements of A are not used in computations). HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of b. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of A. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of A. k >= 0.
[in]APdevice pointer storing the matrix A in banded format.
[in]lda[int] specifies the leading dimension of A. lda >= (k + 1).
[in,out]xdevice pointer storing input vector b. Overwritten by the output vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasDtbsvBatched()

hipblasStatus_t hipblasDtbsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const double *const  AP[],
int  lda,
double *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tbsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a banded triangular matrix, for i = [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A_i*x_i = b_i HIPBLAS_OP_T: Solves A_i**T*x_i = b_i HIPBLAS_OP_C: Solves A_i**H*x_i = b_i
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of each b_i. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of each A_i. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of each A_i. k >= 0.
[in]APdevice vector of device pointers storing each matrix A_i in banded format.
[in]lda[int] specifies the leading dimension of each A_i. lda >= (k + 1).
[in,out]xdevice vector of device pointers storing each input vector b_i. Overwritten by each output vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasDtbsvStridedBatched()

hipblasStatus_t hipblasDtbsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const double *  AP,
int  lda,
hipblasStride  strideA,
double *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tbsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a banded triangular matrix, for i = [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A_i*x_i = b_i HIPBLAS_OP_T: Solves A_i**T*x_i = b_i HIPBLAS_OP_C: Solves A_i**H*x_i = b_i
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of each b_i. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of each A_i. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of each A_i. k >= 0.
[in]APdevice pointer pointing to the first banded matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i. lda >= (k + 1).
[in]strideA[hipblasStride] specifies the distance between the start of one matrix (A_i) and the next (A_i+1).
[in,out]xdevice pointer pointing to the first input vector b_1. Overwritten by output vectors x.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the distance between the start of one vector (x_i) and the next (x_i+1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasDtpmv()

hipblasStatus_t hipblasDtpmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *  AP,
double *  x,
int  incx 
)

BLAS Level 2 API.

tpmv performs one of the matrix-vector operations

 x = A*x or x = A**T*x,

where x is an n element vector and A is an n by n unit, or non-unit, upper or lower triangular matrix, supplied in the pack form.

The vector x is overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of A. m >= 0.
[in]APdevice pointer storing matrix A, of dimension at leat ( m * ( m + 1 ) / 2 ). Before entry with uplo = HIPBLAS_FILL_MODE_UPPER, the array A must contain the upper triangular matrix packed sequentially, column by column, so that A[0] contains a_{0,0}, A[1] and A[2] contain a_{0,1} and a_{1, 1} respectively, and so on. Before entry with uplo = HIPBLAS_FILL_MODE_LOWER, the array A must contain the lower triangular matrix packed sequentially, column by column, so that A[0] contains a_{0,0}, A[1] and A[2] contain a_{1,0} and a_{2,0} respectively, and so on. Note that when DIAG = HIPBLAS_DIAG_UNIT, the diagonal elements of A are not referenced, but are assumed to be unity.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x. incx must not be zero.

◆ hipblasDtpmvBatched()

hipblasStatus_t hipblasDtpmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *const  AP[],
double *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tpmvBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix)

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer storing pointer of matrices A_i, of dimension ( lda, m )
[in]xdevice pointer storing vectors x_i.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasDtpmvStridedBatched()

hipblasStatus_t hipblasDtpmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *  AP,
hipblasStride  strideA,
double *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tpmvStridedBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix) with strides specifying how to retrieve $x_i$ (resp. $A_i$) from $x_{i-1}$ (resp. $A_i$).

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer of the matrix A_0, of dimension ( lda, m )
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_{i + 1}
[in]xdevice pointer storing the vector x_0.
[in]incx[int] specifies the increment for the elements of one vector x.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_{i + 1}
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasDtpsv()

hipblasStatus_t hipblasDtpsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *  AP,
double *  x,
int  incx 
)

BLAS Level 2 API.

tpsv solves

 A*x = b or A**T*x = b, or A**H*x = b,

where x and b are vectors and A is a triangular matrix stored in the packed format.

The input vector b is overwritten by the output vector x.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular (i.e. the diagonal elements of A are not used in computations). HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice pointer storing the packed version of matrix A, of dimension >= (n * (n + 1) / 2)
[in,out]xdevice pointer storing vector b on input, overwritten by x on output.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasDtpsvBatched()

hipblasStatus_t hipblasDtpsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *const  AP[],
double *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tpsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i, or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a triangular matrix stored in the packed format, for i in [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice array of device pointers storing the packed versions of each matrix A_i, of dimension >= (n * (n + 1) / 2)
[in,out]xdevice array of device pointers storing each input vector b_i, overwritten by x_i on output.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasDtpsvStridedBatched()

hipblasStatus_t hipblasDtpsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *  AP,
hipblasStride  strideA,
double *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tpsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i, or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a triangular matrix stored in the packed format, for i in [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice pointer pointing to the first packed matrix A_1, of dimension >= (n * (n + 1) / 2)
[in]strideA[hipblasStride] stride from the beginning of one packed matrix (AP_i) and the next (AP_i+1).
[in,out]xdevice pointer pointing to the first input vector b_1. Overwritten by each x_i on output.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the beginning of one vector (x_i) and the next (x_i+1).
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasDtrmm()

hipblasStatus_t hipblasDtrmm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
double *  BP,
int  ldb 
)

BLAS Level 3 API.

#ifndef HIPBLAS_V1 // deprecated

    hipblasStatus_t hipblasStrmm(hipblasHandle_t    handle,
                                 hipblasSideMode_t  side,
                                 hipblasFillMode_t  uplo,
                                 hipblasOperation_t transA,
                                 hipblasDiagType_t  diag,
                                 int                m,
                                 int                n,
                                 const float*       alpha,
                                 const float*       AP,
                                 int                lda,
                                 float*             BP,
                                 int                ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasStrmmOutofplace(hipblasHandle_t    handle,
                                          hipblasSideMode_t  side,
                                          hipblasFillMode_t  uplo,
                                          hipblasOperation_t transA,
                                          hipblasDiagType_t  diag,
                                          int                m,
                                          int                n,
                                          const float*       alpha,
                                          const float*       AP,
                                          int                lda,
                                          const float*       BP,
                                          int                ldb,
                                          float*             CP,
                                          int                ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasDtrmm(hipblasHandle_t    handle,
                                hipblasSideMode_t  side,
                                hipblasFillMode_t  uplo,
                                hipblasOperation_t transA,
                                hipblasDiagType_t  diag,
                                int                m,
                                int                n,
                                const double*      alpha,
                                const double*      AP,
                                int                lda,
                                double*            BP,
                                int                ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasDtrmmOutofplace(hipblasHandle_t    handle,
                                          hipblasSideMode_t  side,
                                          hipblasFillMode_t  uplo,
                                          hipblasOperation_t transA,
                                          hipblasDiagType_t  diag,
                                          int                m,
                                          int                n,
                                          const double*      alpha,
                                          const double*      AP,
                                          int                lda,
                                          const double*      BP,
                                          int                ldb,
                                          double*            CP,
                                          int                ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasCtrmm(hipblasHandle_t       handle,
                                hipblasSideMode_t     side,
                                hipblasFillMode_t     uplo,
                                hipblasOperation_t    transA,
                                hipblasDiagType_t     diag,
                                int                   m,
                                int                   n,
                                const hipblasComplex* alpha,
                                const hipblasComplex* AP,
                                int                   lda,
                                hipblasComplex*       BP,
                                int                   ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasCtrmmOutofplace(hipblasHandle_t       handle,
                                          hipblasSideMode_t     side,
                                          hipblasFillMode_t     uplo,
                                          hipblasOperation_t    transA,
                                          hipblasDiagType_t     diag,
                                          int                   m,
                                          int                   n,
                                          const hipblasComplex* alpha,
                                          const hipblasComplex* AP,
                                          int                   lda,
                                          const hipblasComplex* BP,
                                          int                   ldb,
                                          hipblasComplex*       CP,
                                          int                   ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasZtrmm(hipblasHandle_t             handle,
                                hipblasSideMode_t           side,
                                hipblasFillMode_t           uplo,
                                hipblasOperation_t          transA,
                                hipblasDiagType_t           diag,
                                int                         m,
                                int                         n,
                                const hipblasDoubleComplex* alpha,
                                const hipblasDoubleComplex* AP,
                                int                         lda,
                                hipblasDoubleComplex*       BP,
                                int                         ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasZtrmmOutofplace(hipblasHandle_t             handle,
                                          hipblasSideMode_t           side,
                                          hipblasFillMode_t           uplo,
                                          hipblasOperation_t          transA,
                                          hipblasDiagType_t           diag,
                                          int                         m,
                                          int                         n,
                                          const hipblasDoubleComplex* alpha,
                                          const hipblasDoubleComplex* AP,
                                          int                         lda,
                                          const hipblasDoubleComplex* BP,
                                          int                         ldb,
                                          hipblasDoubleComplex*       CP,
                                          int                         ldc);

#endif

The deprecated Legacy BLAS in-place trmm performs one of the matrix-matrix operations:

B := alpha*op( A )*B,   or
B := alpha*B*op( A ),

The new trmm performs one of the matrix-matrix operations:

C := alpha*op( A )*B,   or
C := alpha*B*op( A ),

The in-place functionality is still available in the new trmmm by setting pointer C equal to pointer B, and ldc equal to ldb.

alpha  is a scalar,  B  is an m by n matrix, C  is an m by n matrix,  A  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A )  is one  of

op( A ) = A     or
op( A ) = A^T   or
op( A ) = A^H.

When uplo == rocblas_fill_upper the  leading  k by k
upper triangular part of the array  A must contain the upper
triangular matrix and the strictly lower triangular part of
A is not referenced. Here k is m when side == rocblas_side_left
and is n when side == rocblas_side_right.

When uplo == rocblas_fill_lower the  leading  k by k
lower triangular part of the array  A must contain the lower
triangular matrix  and the strictly upper triangular part of
A is not referenced. Here k is m when  side == rocblas_side_left
and is n when side == rocblas_side_right.

Note that when  diag == rocblas_diagonal_unit  the diagonal elements of
A  are not referenced either,  but are assumed to be  unity.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A) multiplies B from the left or right as follows: HIPBLAS_SIDE_LEFT: B := alpha*op( A )*B. HIPBLAS_SIDE_RIGHT: B := alpha*B*op( A ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A is unit triangular as follows: HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B. m >= 0.
[in]n[int] n specifies the number of columns of B. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and B need not be set before entry.
[in]APDevice pointer to matrix A on the GPU. A has dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPDevice pointer to the first matrix B_0 on the GPU. On entry, the leading m by n part of the array B must contain the matrix B, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m ).

◆ hipblasDtrmmBatched()

hipblasStatus_t hipblasDtrmmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const double *  alpha,
const double *const  AP[],
int  lda,
double *const  BP[],
int  ldb,
int  batchCount 
)

BLAS Level 3 API.

The hipBLAS trmm_batched API is from Legacy BLAS and it supports only in-place functionality. It is deprecated and it will be replaced with an API that supports both in-place and out-of-place functionality. The new API is available in hipBLAS versions 1.x.x and later. To get the new API compile with the directive -DHIPBLAS_V1.

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasStrmmBatched(hipblasHandle_t    handle,
                                        hipblasSideMode_t  side,
                                        hipblasFillMode_t  uplo,
                                        hipblasOperation_t transA,
                                        hipblasDiagType_t  diag,
                                        int                m,
                                        int                n,
                                        const float*       alpha,
                                        const float* const AP[],
                                        int                lda,
                                        float* const       BP[],
                                        int                ldb,
                                        int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasStrmmBatchedOutofplace(
                                        hipblasHandle_t    handle,
                                        hipblasSideMode_t  side,
                                        hipblasFillMode_t  uplo,
                                        hipblasOperation_t transA,
                                        hipblasDiagType_t  diag,
                                        int                m,
                                        int                n,
                                        const float*       alpha,
                                        const float* const AP[],
                                        int                lda,
                                        const float* const BP[],
                                        int                ldb,
                                        float* const       CP[],
                                        int                ldc,
                                        int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasDtrmmBatched(hipblasHandle_t     handle,
                                        hipblasSideMode_t   side,
                                        hipblasFillMode_t   uplo,
                                        hipblasOperation_t  transA,
                                        hipblasDiagType_t   diag,
                                        int                 m,
                                        int                 n,
                                        const double*       alpha,
                                        const double* const AP[],
                                        int                 lda,
                                        double* const       BP[],
                                        int                 ldb,
                                        int                 batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasDtrmmBatchedOutofplace(
                                        hipblasHandle_t     handle,
                                        hipblasSideMode_t   side,
                                        hipblasFillMode_t   uplo,
                                        hipblasOperation_t  transA,
                                        hipblasDiagType_t   diag,
                                        int                 m,
                                        int                 n,
                                        const double*       alpha,
                                        const double* const AP[],
                                        int                 lda,
                                        const double* const BP[],
                                        int                 ldb,
                                        double* const       CP[],
                                        int                 ldc,
                                        int                 batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasCtrmmBatched(hipblasHandle_t             handle,
                                        hipblasSideMode_t           side,
                                        hipblasFillMode_t           uplo,
                                        hipblasOperation_t          transA,
                                        hipblasDiagType_t           diag,
                                        int                         m,
                                        int                         n,
                                        const hipblasComplex*       alpha,
                                        const hipblasComplex* const AP[],
                                        int                         lda,
                                        hipblasComplex* const       BP[],
                                        int                         ldb,
                                        int                         batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasCtrmmBatchedOutofplace(
                                        hipblasHandle_t             handle,
                                        hipblasSideMode_t           side,
                                        hipblasFillMode_t           uplo,
                                        hipblasOperation_t          transA,
                                        hipblasDiagType_t           diag,
                                        int                         m,
                                        int                         n,
                                        const hipblasComplex*       alpha,
                                        const hipblasComplex* const AP[],
                                        int                         lda,
                                        const hipblasComplex* const BP[],
                                        int                         ldb,
                                        hipblasComplex* const       CP[],
                                        int                         ldc,
                                        int batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasZtrmmBatched(hipblasHandle_t                   handle,
                                        hipblasSideMode_t                 side,
                                        hipblasFillMode_t                 uplo,
                                        hipblasOperation_t                transA,
                                        hipblasDiagType_t                 diag,
                                        int                               m,
                                        int                               n,
                                        const hipblasDoubleComplex*       alpha,
                                        const hipblasDoubleComplex* const AP[],
                                        int                               lda,
                                        hipblasDoubleComplex* const       BP[],
                                        int                               ldb,
                                        int                               batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasZtrmmBatchedOutofplace(
                                        hipblasHandle_t                   handle,
                                        hipblasSideMode_t                 side,
                                        hipblasFillMode_t                 uplo,
                                        hipblasOperation_t                transA,
                                        hipblasDiagType_t                 diag,
                                        int                               m,
                                        int                               n,
                                        const hipblasDoubleComplex*       alpha,
                                        const hipblasDoubleComplex* const AP[],
                                        int                               lda,
                                        const hipblasDoubleComplex* const BP[],
                                        int                               ldb,
                                        hipblasDoubleComplex* const       CP[],
                                        int                               ldc,
                                        int                               batchCount);

#endif

The deprecated Legacy BLAS in-place trmm_batched performs one of the batched matrix-matrix operations:

B_i := alpha*op( A_i )*B_i,   or
B_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The new trmm_batched performs one of the matrix-matrix operations:

C_i := alpha*op( A_i )*B_i,   or
C_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The in-place functionality is still available in the new trmmm_batched by setting pointer C equal to pointer B and ldc equal to ldb.

alpha  is a scalar,  B_i  is an m by n matrix, C_i  is an m by n matrix,  A_i  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A_i )  is one  of

op( A_i ) = A_i   or   op( A_i ) = A_i^T   or   op( A_i ) = A_i^H.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A_i) multiplies B_i from the left or right as follows: HIPBLAS_SIDE_LEFT: B_i := alpha*op( A_i )*B_i. HIPBLAS_SIDE_RIGHT: B_i := alpha*B_i*op( A_i ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A_i) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A_i) = A_i. HIPBLAS_OP_T: op(A_i) = A_i^T. HIPBLAS_OP_C: op(A_i) = A_i^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A_i is unit triangular as follows: HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i is not referenced and B_i need not be set before entry.
[in]APDevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A_i are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice array of device pointers storing each matrix B_i on the GPU. On entry, the leading m by n part of the array B_i must contain the matrix B_i, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m ).
[in]batchCount[int] number of instances i in the batch.

◆ hipblasDtrmmBatchedOutofplace()

hipblasStatus_t hipblasDtrmmBatchedOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const double *  alpha,
const double *const  AP[],
int  lda,
const double *const  BP[],
int  ldb,
double *const  CP[],
int  ldc,
int  batchCount 
)

◆ hipblasDtrmmOutofplace()

hipblasStatus_t hipblasDtrmmOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
const double *  BP,
int  ldb,
double *  CP,
int  ldc 
)

◆ hipblasDtrmmStridedBatched()

hipblasStatus_t hipblasDtrmmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
double *  BP,
int  ldb,
hipblasStride  strideB,
int  batchCount 
)

BLAS Level 3 API.

The hipBLAS trmm_strided_batched API is from Legacy BLAS and it supports only in-place functionality. It is deprecated and it will be replaced with an API that supports both in-place and out-of-place functionality. The new API is available in hipBLAS versions 1.x.x and later. To get the new API compile with the directive -DHIPBLAS_V1.

#ifndef HIPBLAS_V1  //  deprecated
       hipblasStatus_t hipblasStrmmStridedBatched(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const float*       alpha,
                           const float*       AP,
                           int                lda,
                           hipblasStride      strideA,
                           float*             BP,
                           int                ldb,
                           hipblasStride      strideB,
                           int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasStrmmStridedBatchedOutofplace(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const float*       alpha,
                           const float*       AP,
                           int                lda,
                           hipblasStride      strideA,
                           const float*       BP,
                           int                ldb,
                           hipblasStride      strideB,
                           float*             CP,
                           int                ldc,
                           hipblasStride      strideC,
                           int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated
       hipblasStatus_t hipblasDtrmmStridedBatched(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const double*      alpha,
                           const double*      AP,
                           int                lda,
                           hipblasStride      strideA,
                           double*            BP,
                           int                ldb,
                           hipblasStride      strideB,
                           int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasDtrmmStridedBatchedOutofplace(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const double*      alpha,
                           const double*      AP,
                           int                lda,
                           hipblasStride      strideA,
                           const double*      BP,
                           int                ldb,
                           hipblasStride      strideB,
                           double*            CP,
                           int                ldc,
                           hipblasStride      strideC,
                           int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

       hipblasStatus_t hipblasCtrmmStridedBatched(
                           hipblasHandle_t       handle,
                           hipblasSideMode_t     side,
                           hipblasFillMode_t     uplo,
                           hipblasOperation_t    transA,
                           hipblasDiagType_t     diag,
                           int                   m,
                           int                   n,
                           const hipblasComplex* alpha,
                           const hipblasComplex* AP,
                           int                   lda,
                           hipblasStride         strideA,
                           hipblasComplex*       BP,
                           int                   ldb,
                           hipblasStride         strideB,
                           int                   batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasCtrmmStridedBatchedOutofplace(
                           hipblasHandle_t       handle,
                           hipblasSideMode_t     side,
                           hipblasFillMode_t     uplo,
                           hipblasOperation_t    transA,
                           hipblasDiagType_t     diag,
                           int                   m,
                           int                   n,
                           const hipblasComplex* alpha,
                           const hipblasComplex* AP,
                           int                   lda,
                           hipblasStride         strideA,
                           const hipblasComplex* BP,
                           int                   ldb,
                           hipblasStride         strideB,
                           hipblasComplex*       CP,
                           int                   ldc,
                           hipblasStride         strideC,
                           int batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasZtrmmStridedBatched(hipblasHandle_t             handle,
                                               hipblasSideMode_t           side,
                                               hipblasFillMode_t           uplo,
                                               hipblasOperation_t          transA,
                                               hipblasDiagType_t           diag,
                                               int                         m,
                                               int                         n,
                                               const hipblasDoubleComplex* alpha,
                                               const hipblasDoubleComplex* AP,
                                               int                         lda,
                                               hipblasStride               strideA,
                                               hipblasDoubleComplex*       BP,
                                               int                         ldb,
                                               hipblasStride               strideB,
                                               int                         batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasZtrmmStridedBatchedOutofplace(
                                                hipblasHandle_t             handle,
                                                hipblasSideMode_t           side,
                                                hipblasFillMode_t           uplo,
                                                hipblasOperation_t          transA,
                                                hipblasDiagType_t           diag,
                                                int                         m,
                                                int                         n,
                                                const hipblasDoubleComplex* alpha,
                                                const hipblasDoubleComplex* AP,
                                                int                         lda,
                                                hipblasStride               strideA,
                                                const hipblasDoubleComplex* BP,
                                                int                         ldb,
                                                hipblasStride               strideB,
                                                hipblasDoubleComplex*       BC,
                                                int                         ldc,
                                                hipblasStride               strideC,
                                                int                         batchCount);

#endif

The deprecated Legacy BLAS in-place trmm_strided_batched performs one of the strided_batched matrix-matrix operations:

B_i := alpha*op( A_i )*B_i,   or
B_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The new trmm_batched performs one of the matrix-matrix operations:

C_i := alpha*op( A_i )*B_i,   or
C_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The in-place functionality is still available in the new trmmm_batched by setting pointer C equal to pointer B, setting ldc equal to ldb, and setting stride_C equal to stride_B.

alpha  is a scalar,  B_i  is an m by n matrix, C_i  is an m by n matrix,  A_i  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A_i )  is one  of

op( A_i ) = A_i   or
op( A_i ) = A_i^T   or
op( A_i ) = A_i^H.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A_i) multiplies B_i from the left or right as follows: HIPBLAS_SIDE_LEFT: B_i := alpha*op( A_i )*B_i. HIPBLAS_SIDE_RIGHT: B_i := alpha*B_i*op( A_i ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A_i) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A_i) = A_i. HIPBLAS_OP_T: op(A_i) = A_i^T. HIPBLAS_OP_C: op(A_i) = A_i^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A_i is unit triangular as follows: HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i is not referenced and B_i need not be set before entry.
[in]APDevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A_i are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in,out]BPDevice pointer to the first matrix B_0 on the GPU. On entry, the leading m by n part of the array B_i must contain the matrix B_i, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasDtrmmStridedBatchedOutofplace()

hipblasStatus_t hipblasDtrmmStridedBatchedOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
const double *  BP,
int  ldb,
hipblasStride  strideB,
double *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

◆ hipblasDtrmv()

hipblasStatus_t hipblasDtrmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *  AP,
int  lda,
double *  x,
int  incx 
)

BLAS Level 2 API.

trmv performs one of the matrix-vector operations

 x = A*x or x = A**T*x,

where x is an n element vector and A is an n by n unit, or non-unit, upper or lower triangular matrix.

The vector x is overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of A. m >= 0.
[in]APdevice pointer storing matrix A, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A. lda = max( 1, m ).
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasDtrmvBatched()

hipblasStatus_t hipblasDtrmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *const  AP[],
int  lda,
double *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

trmvBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix)

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer storing pointer of matrices A_i, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A_i. lda >= max( 1, m ).
[in]xdevice pointer storing vectors x_i.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasDtrmvStridedBatched()

hipblasStatus_t hipblasDtrmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *  AP,
int  lda,
hipblasStride  strideA,
double *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

trmvStridedBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix) with strides specifying how to retrieve $x_i$ (resp. $A_i$) from $x_{i-1}$ (resp. $A_i$).

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer of the matrix A_0, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A_i. lda >= max( 1, m ).
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_{i + 1}
[in]xdevice pointer storing the vector x_0.
[in]incx[int] specifies the increment for the elements of one vector x.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_{i + 1}
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasDtrsm()

hipblasStatus_t hipblasDtrsm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
double *  BP,
int  ldb 
)

BLAS Level 3 API.

trsm solves

op(A)*X = alpha*B or  X*op(A) = alpha*B,

where alpha is a scalar, X and B are m by n matrices, A is triangular matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

The matrix X is overwritten on B.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time.

(where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B. m >= 0.
[in]n[int] n specifies the number of columns of B. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice pointer storing matrix A. of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice pointer storing matrix B.
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m ).

◆ hipblasDtrsmBatched()

hipblasStatus_t hipblasDtrsmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const double *  alpha,
const double *const  AP[],
int  lda,
double *const  BP[],
int  ldb,
int  batchCount 
)

BLAS Level 3 API.

trsmBatched performs the following batched operation:

op(A_i)*X_i = alpha*B_i or  X_i*op(A_i) = alpha*B_i, for i = 1, ..., batchCount.

where alpha is a scalar, X and B are batched m by n matrices, A is triangular batched matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

Each matrix X_i is overwritten on B_i for i = 1, ..., batchCount.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time. (where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each B_i. m >= 0.
[in]n[int] n specifies the number of columns of each B_i. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Matricies are of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of each A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice array of device pointers storing each matrix B_i on the GPU.
[in]ldb[int] ldb specifies the first dimension of each B_i. ldb >= max( 1, m ).
[in]batchCount[int] number of trsm operatons in the batch.

◆ hipblasDtrsmStridedBatched()

hipblasStatus_t hipblasDtrsmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const double *  alpha,
const double *  AP,
int  lda,
hipblasStride  strideA,
double *  BP,
int  ldb,
hipblasStride  strideB,
int  batchCount 
)

BLAS Level 3 API.

trsmSridedBatched performs the following strided batched operation:

op(A_i)*X_i = alpha*B_i or  X_i*op(A_i) = alpha*B_i, for i = 1, ..., batchCount.

where alpha is a scalar, X and B are strided batched m by n matrices, A is triangular strided batched matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

Each matrix X_i is overwritten on B_i for i = 1, ..., batchCount.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time. (where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each B_i. m >= 0.
[in]n[int] n specifies the number of columns of each B_i. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice pointer pointing to the first matrix A_1. of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of each A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in,out]BPdevice pointer pointing to the first matrix B_1.
[in]ldb[int] ldb specifies the first dimension of each B_i. ldb >= max( 1, m ).
[in]strideB[hipblasStride] stride from the start of one B_i matrix to the next B_(i + 1).
[in]batchCount[int] number of trsm operatons in the batch.

◆ hipblasDtrsv()

hipblasStatus_t hipblasDtrsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *  AP,
int  lda,
double *  x,
int  incx 
)

BLAS Level 2 API.

trsv solves

 A*x = b or A**T*x = b,

where x and b are vectors and A is a triangular matrix.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice pointer storing matrix A, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A. lda = max( 1, m ).
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasDtrsvBatched()

hipblasStatus_t hipblasDtrsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *const  AP[],
int  lda,
double *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

trsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i,

where (A_i, x_i, b_i) is the i-th instance of the batch. x_i and b_i are vectors and A_i is an m by m triangular matrix.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i. lda = max(1, m)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of x.
[in]batchCount[int] number of instances in the batch

◆ hipblasDtrsvStridedBatched()

hipblasStatus_t hipblasDtrsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const double *  AP,
int  lda,
hipblasStride  strideA,
double *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

trsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i,

where (A_i, x_i, b_i) is the i-th instance of the batch. x_i and b_i are vectors and A_i is an m by m triangular matrix, for i = 1, ..., batchCount.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice pointer to the first matrix (A_1) in the batch, of dimension ( lda, m )
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1)
[in]lda[int] specifies the leading dimension of each A_i. lda = max( 1, m ).
[in,out]xdevice pointer to the first vector (x_1) in the batch.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_(i + 1)
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasDtrtri()

hipblasStatus_t hipblasDtrtri ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const double *  AP,
int  lda,
double *  invA,
int  ldinvA 
)

BLAS Level 3 API.

trtri compute the inverse of a matrix A, namely, invA

and write the result into invA;
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int] size of matrix A and invA
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[out]invAdevice pointer storing matrix invA.
[in]ldinvA[int] specifies the leading dimension of invA.

◆ hipblasDtrtriBatched()

hipblasStatus_t hipblasDtrtriBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const double *const  AP[],
int  lda,
double *  invA[],
int  ldinvA,
int  batchCount 
)

BLAS Level 3 API.

trtriBatched compute the inverse of A_i and write into invA_i where A_i and invA_i are the i-th matrices in the batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER'
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int]
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[out]invAdevice array of device pointers storing the inverse of each matrix A_i. Partial inplace operation is supported, see below. If UPLO = 'U', the leading N-by-N upper triangular part of the invA will store the inverse of the upper triangular matrix, and the strictly lower triangular part of invA is cleared. If UPLO = 'L', the leading N-by-N lower triangular part of the invA will store the inverse of the lower triangular matrix, and the strictly upper triangular part of invA is cleared.
[in]ldinvA[int] specifies the leading dimension of each invA_i.
[in]batchCount[int] numbers of matrices in the batch

◆ hipblasDtrtriStridedBatched()

hipblasStatus_t hipblasDtrtriStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const double *  AP,
int  lda,
hipblasStride  strideA,
double *  invA,
int  ldinvA,
hipblasStride  stride_invA,
int  batchCount 
)

BLAS Level 3 API.

trtriStridedBatched compute the inverse of A_i and write into invA_i where A_i and invA_i are the i-th matrices in the batch, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER'
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int]
[in]APdevice pointer pointing to address of first matrix A_1.
[in]lda[int] specifies the leading dimension of each A.
[in]strideA[hipblasStride] "batch stride a": stride from the start of one A_i matrix to the next A_(i + 1).
[out]invAdevice pointer storing the inverses of each matrix A_i. Partial inplace operation is supported, see below. If UPLO = 'U', the leading N-by-N upper triangular part of the invA will store the inverse of the upper triangular matrix, and the strictly lower triangular part of invA is cleared. If UPLO = 'L', the leading N-by-N lower triangular part of the invA will store the inverse of the lower triangular matrix, and the strictly upper triangular part of invA is cleared.
[in]ldinvA[int] specifies the leading dimension of each invA_i.
[in]stride_invA[hipblasStride] "batch stride invA": stride from the start of one invA_i matrix to the next invA_(i + 1).
[in]batchCount[int] numbers of matrices in the batch

◆ hipblasDzasum()

hipblasStatus_t hipblasDzasum ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
double *  result 
)

BLAS Level 1 API.

asum computes the sum of the magnitudes of elements of a real vector x, or the sum of magnitudes of the real and imaginary parts of elements if x is a complex vector.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x. incx must be > 0.
[in,out]resultdevice pointer or host pointer to store the asum product. return is 0.0 if n <= 0.

◆ hipblasDzasumBatched()

hipblasStatus_t hipblasDzasumBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount,
double *  result 
)

BLAS Level 1 API.

asumBatched computes the sum of the magnitudes of the elements in a batch of real vectors x_i, or the sum of magnitudes of the real and imaginary parts of elements if x_i is a complex vector, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch.
[out]resultdevice array or host array of batchCount size for results. return is 0.0 if n, incx<=0.

◆ hipblasDzasumStridedBatched()

hipblasStatus_t hipblasDzasumStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount,
double *  result 
)

BLAS Level 1 API.

asumStridedBatched computes the sum of the magnitudes of elements of a real vectors x_i, or the sum of magnitudes of the real and imaginary parts of elements if x_i is a complex vector, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array for storing contiguous batchCount results. return is 0.0 if n, incx<=0.

◆ hipblasDznrm2()

hipblasStatus_t hipblasDznrm2 ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
double *  result 
)

BLAS Level 1 API.

nrm2 computes the euclidean norm of a real or complex vector

      result := sqrt( x'*x ) for real vectors
      result := sqrt( x**H*x ) for complex vectors
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : s,d,sc,dz
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the nrm2 product. return is 0.0 if n, incx<=0.

◆ hipblasDznrm2Batched()

hipblasStatus_t hipblasDznrm2Batched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount,
double *  result 
)

BLAS Level 1 API.

nrm2Batched computes the euclidean norm over a batch of real or complex vectors

      result := sqrt( x_i'*x_i ) for real vectors x, for i = 1, ..., batchCount
      result := sqrt( x_i**H*x_i ) for complex vectors x, for i = 1, ..., batchCount
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array of batchCount size for nrm2 results. return is 0.0 for each element if n <= 0, incx<=0.

◆ hipblasDznrm2StridedBatched()

hipblasStatus_t hipblasDznrm2StridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount,
double *  result 
)

BLAS Level 1 API.

nrm2StridedBatched computes the euclidean norm over a batch of real or complex vectors

      := sqrt( x_i'*x_i ) for real vectors x, for i = 1, ..., batchCount
      := sqrt( x_i**H*x_i ) for complex vectors, for i = 1, ..., batchCount
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array for storing contiguous batchCount results. return is 0.0 for each element if n <= 0, incx<=0.

◆ hipblasGemmBatchedEx()

hipblasStatus_t hipblasGemmBatchedEx ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const void *  alpha,
const void *  A[],
hipblasDatatype_t  aType,
int  lda,
const void *  B[],
hipblasDatatype_t  bType,
int  ldb,
const void *  beta,
void *  C[],
hipblasDatatype_t  cType,
int  ldc,
int  batchCount,
hipblasDatatype_t  computeType,
hipblasGemmAlgo_t  algo 
)

BLAS EX API.

gemmBatchedEx performs one of the batched matrix-matrix operations C_i = alpha*op(A_i)*op(B_i) + beta*C_i, for i = 1, ..., batchCount. where op( X ) is one of op( X ) = X or op( X ) = X**T or op( X ) = X**H, alpha and beta are scalars, and A, B, and C are batched pointers to matrices, with op( A ) an m by k by batchCount batched matrix, op( B ) a k by n by batchCount batched matrix and C a m by n by batchCount batched matrix. The batched matrices are an array of pointers to matrices. The number of pointers to matrices is batchCount.

  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A ).
[in]transB[hipblasOperation_t] specifies the form of op( B ).
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]k[int] matrix dimension k.
[in]alpha[const void *] device pointer or host pointer specifying the scalar alpha. Same datatype as computeType.
[in]A[void *] device pointer storing array of pointers to each matrix A_i.
[in]aType[hipblasDatatype_t] specifies the datatype of each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]B[void *] device pointer storing array of pointers to each matrix B_i.
[in]bType[hipblasDatatype_t] specifies the datatype of each matrix B_i.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]beta[const void *] device pointer or host pointer specifying the scalar beta. Same datatype as computeType.
[in]C[void *] device array of device pointers to each matrix C_i.
[in]cType[hipblasDatatype_t] specifies the datatype of each matrix C_i.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]batchCount[int] number of gemm operations in the batch.
[in]computeType[hipblasDatatype_t] specifies the datatype of computation.
[in]algo[hipblasGemmAlgo_t] enumerant specifying the algorithm type.

◆ hipblasGemmEx()

hipblasStatus_t hipblasGemmEx ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const void *  alpha,
const void *  A,
hipblasDatatype_t  aType,
int  lda,
const void *  B,
hipblasDatatype_t  bType,
int  ldb,
const void *  beta,
void *  C,
hipblasDatatype_t  cType,
int  ldc,
hipblasDatatype_t  computeType,
hipblasGemmAlgo_t  algo 
)

BLAS EX API.

gemmEx performs one of the matrix-matrix operations

C = alpha*op( A )*op( B ) + beta*C,

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B, and C are matrices, with op( A ) an m by k matrix, op( B ) a k by n matrix and C is a m by n matrix.

  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A ).
[in]transB[hipblasOperation_t] specifies the form of op( B ).
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]k[int] matrix dimension k.
[in]alpha[const void *] device pointer or host pointer specifying the scalar alpha. Same datatype as computeType.
[in]A[void *] device pointer storing matrix A.
[in]aType[hipblasDatatype_t] specifies the datatype of matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]B[void *] device pointer storing matrix B.
[in]bType[hipblasDatatype_t] specifies the datatype of matrix B.
[in]ldb[int] specifies the leading dimension of B.
[in]beta[const void *] device pointer or host pointer specifying the scalar beta. Same datatype as computeType.
[in]C[void *] device pointer storing matrix C.
[in]cType[hipblasDatatype_t] specifies the datatype of matrix C.
[in]ldc[int] specifies the leading dimension of C.
[in]computeType[hipblasDatatype_t] specifies the datatype of computation.
[in]algo[hipblasGemmAlgo_t] enumerant specifying the algorithm type.

◆ hipblasGemmStridedBatchedEx()

hipblasStatus_t hipblasGemmStridedBatchedEx ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const void *  alpha,
const void *  A,
hipblasDatatype_t  aType,
int  lda,
hipblasStride  strideA,
const void *  B,
hipblasDatatype_t  bType,
int  ldb,
hipblasStride  strideB,
const void *  beta,
void *  C,
hipblasDatatype_t  cType,
int  ldc,
hipblasStride  strideC,
int  batchCount,
hipblasDatatype_t  computeType,
hipblasGemmAlgo_t  algo 
)

BLAS EX API.

gemmStridedBatchedEx performs one of the strided_batched matrix-matrix operations

C_i = alpha*op(A_i)*op(B_i) + beta*C_i, for i = 1, ..., batchCount

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B, and C are strided_batched matrices, with op( A ) an m by k by batchCount strided_batched matrix, op( B ) a k by n by batchCount strided_batched matrix and C a m by n by batchCount strided_batched matrix.

The strided_batched matrices are multiple matrices separated by a constant stride. The number of matrices is batchCount.

  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A ).
[in]transB[hipblasOperation_t] specifies the form of op( B ).
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]k[int] matrix dimension k.
[in]alpha[const void *] device pointer or host pointer specifying the scalar alpha. Same datatype as computeType.
[in]A[void *] device pointer pointing to first matrix A_1.
[in]aType[hipblasDatatype_t] specifies the datatype of each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] specifies stride from start of one A_i matrix to the next A_(i + 1).
[in]B[void *] device pointer pointing to first matrix B_1.
[in]bType[hipblasDatatype_t] specifies the datatype of each matrix B_i.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]strideB[hipblasStride] specifies stride from start of one B_i matrix to the next B_(i + 1).
[in]beta[const void *] device pointer or host pointer specifying the scalar beta. Same datatype as computeType.
[in]C[void *] device pointer pointing to first matrix C_1.
[in]cType[hipblasDatatype_t] specifies the datatype of each matrix C_i.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]strideC[hipblasStride] specifies stride from start of one C_i matrix to the next C_(i + 1).
[in]batchCount[int] number of gemm operations in the batch.
[in]computeType[hipblasDatatype_t] specifies the datatype of computation.
[in]algo[hipblasGemmAlgo_t] enumerant specifying the algorithm type.

◆ hipblasGetAtomicsMode()

hipblasStatus_t hipblasGetAtomicsMode ( hipblasHandle_t  handle,
hipblasAtomicsMode_t atomics_mode 
)

Get hipblasSetAtomicsMode.

◆ hipblasGetMatrix()

hipblasStatus_t hipblasGetMatrix ( int  rows,
int  cols,
int  elemSize,
const void *  AP,
int  lda,
void *  BP,
int  ldb 
)

copy matrix from device to host

Parameters
[in]rows[int] number of rows in matrices
[in]cols[int] number of columns in matrices
[in]elemSize[int] number of bytes per element in the matrix
[in]APpointer to matrix on the GPU
[in]lda[int] specifies the leading dimension of A, lda >= rows
[out]BPpointer to matrix on the host
[in]ldb[int] specifies the leading dimension of B, ldb >= rows

◆ hipblasGetMatrixAsync()

hipblasStatus_t hipblasGetMatrixAsync ( int  rows,
int  cols,
int  elemSize,
const void *  AP,
int  lda,
void *  BP,
int  ldb,
hipStream_t  stream 
)

asynchronously copy matrix from device to host

hipblasGetMatrixAsync copies a matrix from device memory to pinned host memory asynchronously. Memory on the host must be allocated with hipHostMalloc or the transfer will be synchronous.

Parameters
[in]rows[int] number of rows in matrices
[in]cols[int] number of columns in matrices
[in]elemSize[int] number of bytes per element in the matrix
[in]APpointer to matrix on the GPU
[in]lda[int] specifies the leading dimension of A, lda >= rows
[out]BPpointer to matrix on the host
[in]ldb[int] specifies the leading dimension of B, ldb >= rows
[in]streamspecifies the stream into which this transfer request is queued

◆ hipblasGetPointerMode()

hipblasStatus_t hipblasGetPointerMode ( hipblasHandle_t  handle,
hipblasPointerMode_t mode 
)

Get hipblas pointer mode.

◆ hipblasGetStream()

hipblasStatus_t hipblasGetStream ( hipblasHandle_t  handle,
hipStream_t *  streamId 
)

Get stream[0] for handle.

◆ hipblasGetVector()

hipblasStatus_t hipblasGetVector ( int  n,
int  elemSize,
const void *  x,
int  incx,
void *  y,
int  incy 
)

copy vector from device to host

Parameters
[in]n[int] number of elements in the vector
[in]elemSize[int] Size of both vectors in bytes
[in]xpointer to vector on the device
[in]incx[int] specifies the increment for the elements of the vector
[out]ypointer to vector on the host
[in]incy[int] specifies the increment for the elements of the vector

◆ hipblasGetVectorAsync()

hipblasStatus_t hipblasGetVectorAsync ( int  n,
int  elemSize,
const void *  x,
int  incx,
void *  y,
int  incy,
hipStream_t  stream 
)

asynchronously copy vector from device to host

hipblasGetVectorAsync copies a vector from pinned host memory to device memory asynchronously. Memory on the host must be allocated with hipHostMalloc or the transfer will be synchronous.

Parameters
[in]n[int] number of elements in the vector
[in]elemSize[int] number of bytes per element in the matrix
[in]xpointer to vector on the device
[in]incx[int] specifies the increment for the elements of the vector
[out]ypointer to vector on the host
[in]incy[int] specifies the increment for the elements of the vector
[in]streamspecifies the stream into which this transfer request is queued

◆ hipblasHaxpy()

hipblasStatus_t hipblasHaxpy ( hipblasHandle_t  handle,
int  n,
const hipblasHalf alpha,
const hipblasHalf x,
int  incx,
hipblasHalf y,
int  incy 
)

BLAS Level 1 API.

axpy computes constant alpha multiplied by vector x, plus vector y

y := alpha * x + y
  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]alphadevice pointer or host pointer to specify the scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[out]ydevice pointer storing vector y.
[in,out]incy[int] specifies the increment for the elements of y.

◆ hipblasHaxpyBatched()

hipblasStatus_t hipblasHaxpyBatched ( hipblasHandle_t  handle,
int  n,
const hipblasHalf alpha,
const hipblasHalf *const  x[],
int  incx,
hipblasHalf *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

axpyBatched compute y := alpha * x + y over a set of batched vectors.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]alphaspecifies the scalar alpha.
[in]xpointer storing vector x on the GPU.
[in]incx[int] specifies the increment for the elements of x.
[out]ypointer storing vector y on the GPU.
[in,out]incy[int] specifies the increment for the elements of y.
[in]batchCount[int] number of instances in the batch

◆ hipblasHaxpyStridedBatched()

hipblasStatus_t hipblasHaxpyStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasHalf alpha,
const hipblasHalf x,
int  incx,
hipblasStride  stridex,
hipblasHalf y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

axpyStridedBatched compute y := alpha * x + y over a set of strided batched vectors.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int]
[in]alphaspecifies the scalar alpha.
[in]xpointer storing vector x on the GPU.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] specifies the increment between vectors of x.
[out]ypointer storing vector y on the GPU.
[in,out]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] specifies the increment between vectors of y.
[in]batchCount[int] number of instances in the batch

◆ hipblasHdot()

hipblasStatus_t hipblasHdot ( hipblasHandle_t  handle,
int  n,
const hipblasHalf x,
int  incx,
const hipblasHalf y,
int  incy,
hipblasHalf result 
)

BLAS Level 1 API.

dot(u) performs the dot product of vectors x and y

result = x * y;

dotc performs the dot product of the conjugate of complex vector x and complex vector y

result = conjugate (x) * y;
  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the dot product. return is 0.0 if n <= 0.

◆ hipblasHdotBatched()

hipblasStatus_t hipblasHdotBatched ( hipblasHandle_t  handle,
int  n,
const hipblasHalf *const  x[],
int  incx,
const hipblasHalf *const  y[],
int  incy,
int  batchCount,
hipblasHalf result 
)

BLAS Level 1 API.

dotBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasHdotStridedBatched()

hipblasStatus_t hipblasHdotStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasHalf x,
int  incx,
hipblasStride  stridex,
const hipblasHalf y,
int  incy,
hipblasStride  stridey,
int  batchCount,
hipblasHalf result 
)

BLAS Level 1 API.

dotStridedBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcStridedBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasHgemm()

hipblasStatus_t hipblasHgemm ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const hipblasHalf alpha,
const hipblasHalf AP,
int  lda,
const hipblasHalf BP,
int  ldb,
const hipblasHalf beta,
hipblasHalf CP,
int  ldc 
)

BLAS Level 3 API.

gemm performs one of the matrix-matrix operations

C = alpha*op( A )*op( B ) + beta*C,

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are matrices, with op( A ) an m by k matrix, op( B ) a k by n matrix and C an m by n matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t]
      .
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] number or rows of matrices op( A ) and C
[in]n[int] number of columns of matrices op( B ) and C
[in]k[int] number of columns of matrix op( A ) and number of rows of matrix op( B )
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]BPdevice pointer storing matrix B.
[in]ldb[int] specifies the leading dimension of B.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice pointer storing matrix C on the GPU.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasHgemmBatched()

hipblasStatus_t hipblasHgemmBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const hipblasHalf alpha,
const hipblasHalf *const  AP[],
int  lda,
const hipblasHalf *const  BP[],
int  ldb,
const hipblasHalf beta,
hipblasHalf *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

gemmBatched performs one of the batched matrix-matrix operations C_i = alpha*op( A_i )*op( B_i ) + beta*C_i, for i = 1, ..., batchCount. where op( X ) is one of op( X ) = X or op( X ) = X**T or op( X ) = X**H, alpha and beta are scalars, and A, B and C are strided batched matrices, with op( A ) an m by k by batchCount strided_batched matrix, op( B ) an k by n by batchCount strided_batched matrix and C an m by n by batchCount strided_batched matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimention m.
[in]n[int] matrix dimention n.
[in]k[int] matrix dimention k.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]BPdevice array of device pointers storing each matrix B_i.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice array of device pointers storing each matrix C_i.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]batchCount[int] number of gemm operations in the batch

◆ hipblasHgemmStridedBatched()

hipblasStatus_t hipblasHgemmStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const hipblasHalf alpha,
const hipblasHalf AP,
int  lda,
long long  strideA,
const hipblasHalf BP,
int  ldb,
long long  strideB,
const hipblasHalf beta,
hipblasHalf CP,
int  ldc,
long long  strideC,
int  batchCount 
)

BLAS Level 3 API.

gemmStridedBatched performs one of the strided batched matrix-matrix operations

C_i = alpha*op( A_i )*op( B_i ) + beta*C_i, for i = 1, ..., batchCount.

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are strided batched matrices, with op( A ) an m by k by batchCount strided_batched matrix, op( B ) an k by n by batchCount strided_batched matrix and C an m by n by batchCount strided_batched matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimention m.
[in]n[int] matrix dimention n.
[in]k[int] matrix dimention k.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer pointing to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in]BPdevice pointer pointing to the first matrix B_1.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]strideB[hipblasStride] stride from the start of one B_i matrix to the next B_(i + 1).
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice pointer pointing to the first matrix C_1.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]strideC[hipblasStride] stride from the start of one C_i matrix to the next C_(i + 1).
[in]batchCount[int] number of gemm operatons in the batch

◆ hipblasIcamax()

hipblasStatus_t hipblasIcamax ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
int *  result 
)

BLAS Level 1 API.

amax finds the first index of the element of maximum magnitude of a vector x.

  • Supported precisions in rocBLAS : s,d,c,z.
  • Supported precisions in cuBLAS : s,d,c,z.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the amax index. return is 0.0 if n, incx<=0.

◆ hipblasIcamaxBatched()

hipblasStatus_t hipblasIcamaxBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex *const  x[],
int  incx,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

amaxBatched finds the first index of the element of maximum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z.
  • Supported precisions in cuBLAS : No support.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch, must be > 0.
[out]resultdevice or host array of pointers of batchCount size for results. return is 0 if n, incx<=0.

◆ hipblasIcamaxStridedBatched()

hipblasStatus_t hipblasIcamaxStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

amaxStridedBatched finds the first index of the element of maximum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] specifies the pointer increment between one x_i and the next x_(i + 1).
[in]batchCount[int] number of instances in the batch
[out]resultdevice or host pointer for storing contiguous batchCount results. return is 0 if n <= 0, incx<=0.

◆ hipblasIcamin()

hipblasStatus_t hipblasIcamin ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
int *  result 
)

BLAS Level 1 API.

amin finds the first index of the element of minimum magnitude of a vector x.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the amin index. return is 0.0 if n, incx<=0.

◆ hipblasIcaminBatched()

hipblasStatus_t hipblasIcaminBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex *const  x[],
int  incx,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

aminBatched finds the first index of the element of minimum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch, must be > 0.
[out]resultdevice or host pointers to array of batchCount size for results. return is 0 if n, incx<=0.

◆ hipblasIcaminStridedBatched()

hipblasStatus_t hipblasIcaminStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

aminStridedBatched finds the first index of the element of minimum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] specifies the pointer increment between one x_i and the next x_(i + 1)
[in]batchCount[int] number of instances in the batch
[out]resultdevice or host pointer to array for storing contiguous batchCount results. return is 0 if n <= 0, incx<=0.

◆ hipblasIdamax()

hipblasStatus_t hipblasIdamax ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
int *  result 
)

BLAS Level 1 API.

amax finds the first index of the element of maximum magnitude of a vector x.

  • Supported precisions in rocBLAS : s,d,c,z.
  • Supported precisions in cuBLAS : s,d,c,z.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the amax index. return is 0.0 if n, incx<=0.

◆ hipblasIdamaxBatched()

hipblasStatus_t hipblasIdamaxBatched ( hipblasHandle_t  handle,
int  n,
const double *const  x[],
int  incx,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

amaxBatched finds the first index of the element of maximum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z.
  • Supported precisions in cuBLAS : No support.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch, must be > 0.
[out]resultdevice or host array of pointers of batchCount size for results. return is 0 if n, incx<=0.

◆ hipblasIdamaxStridedBatched()

hipblasStatus_t hipblasIdamaxStridedBatched ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
hipblasStride  stridex,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

amaxStridedBatched finds the first index of the element of maximum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] specifies the pointer increment between one x_i and the next x_(i + 1).
[in]batchCount[int] number of instances in the batch
[out]resultdevice or host pointer for storing contiguous batchCount results. return is 0 if n <= 0, incx<=0.

◆ hipblasIdamin()

hipblasStatus_t hipblasIdamin ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
int *  result 
)

BLAS Level 1 API.

amin finds the first index of the element of minimum magnitude of a vector x.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the amin index. return is 0.0 if n, incx<=0.

◆ hipblasIdaminBatched()

hipblasStatus_t hipblasIdaminBatched ( hipblasHandle_t  handle,
int  n,
const double *const  x[],
int  incx,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

aminBatched finds the first index of the element of minimum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch, must be > 0.
[out]resultdevice or host pointers to array of batchCount size for results. return is 0 if n, incx<=0.

◆ hipblasIdaminStridedBatched()

hipblasStatus_t hipblasIdaminStridedBatched ( hipblasHandle_t  handle,
int  n,
const double *  x,
int  incx,
hipblasStride  stridex,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

aminStridedBatched finds the first index of the element of minimum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] specifies the pointer increment between one x_i and the next x_(i + 1)
[in]batchCount[int] number of instances in the batch
[out]resultdevice or host pointer to array for storing contiguous batchCount results. return is 0 if n <= 0, incx<=0.

◆ hipblasIsamax()

hipblasStatus_t hipblasIsamax ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
int *  result 
)

BLAS Level 1 API.

amax finds the first index of the element of maximum magnitude of a vector x.

  • Supported precisions in rocBLAS : s,d,c,z.
  • Supported precisions in cuBLAS : s,d,c,z.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the amax index. return is 0.0 if n, incx<=0.

◆ hipblasIsamaxBatched()

hipblasStatus_t hipblasIsamaxBatched ( hipblasHandle_t  handle,
int  n,
const float *const  x[],
int  incx,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

amaxBatched finds the first index of the element of maximum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z.
  • Supported precisions in cuBLAS : No support.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch, must be > 0.
[out]resultdevice or host array of pointers of batchCount size for results. return is 0 if n, incx<=0.

◆ hipblasIsamaxStridedBatched()

hipblasStatus_t hipblasIsamaxStridedBatched ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
hipblasStride  stridex,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

amaxStridedBatched finds the first index of the element of maximum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] specifies the pointer increment between one x_i and the next x_(i + 1).
[in]batchCount[int] number of instances in the batch
[out]resultdevice or host pointer for storing contiguous batchCount results. return is 0 if n <= 0, incx<=0.

◆ hipblasIsamin()

hipblasStatus_t hipblasIsamin ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
int *  result 
)

BLAS Level 1 API.

amin finds the first index of the element of minimum magnitude of a vector x.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the amin index. return is 0.0 if n, incx<=0.

◆ hipblasIsaminBatched()

hipblasStatus_t hipblasIsaminBatched ( hipblasHandle_t  handle,
int  n,
const float *const  x[],
int  incx,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

aminBatched finds the first index of the element of minimum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch, must be > 0.
[out]resultdevice or host pointers to array of batchCount size for results. return is 0 if n, incx<=0.

◆ hipblasIsaminStridedBatched()

hipblasStatus_t hipblasIsaminStridedBatched ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
hipblasStride  stridex,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

aminStridedBatched finds the first index of the element of minimum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] specifies the pointer increment between one x_i and the next x_(i + 1)
[in]batchCount[int] number of instances in the batch
[out]resultdevice or host pointer to array for storing contiguous batchCount results. return is 0 if n <= 0, incx<=0.

◆ hipblasIzamax()

hipblasStatus_t hipblasIzamax ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
int *  result 
)

BLAS Level 1 API.

amax finds the first index of the element of maximum magnitude of a vector x.

  • Supported precisions in rocBLAS : s,d,c,z.
  • Supported precisions in cuBLAS : s,d,c,z.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the amax index. return is 0.0 if n, incx<=0.

◆ hipblasIzamaxBatched()

hipblasStatus_t hipblasIzamaxBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

amaxBatched finds the first index of the element of maximum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z.
  • Supported precisions in cuBLAS : No support.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch, must be > 0.
[out]resultdevice or host array of pointers of batchCount size for results. return is 0 if n, incx<=0.

◆ hipblasIzamaxStridedBatched()

hipblasStatus_t hipblasIzamaxStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

amaxStridedBatched finds the first index of the element of maximum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] specifies the pointer increment between one x_i and the next x_(i + 1).
[in]batchCount[int] number of instances in the batch
[out]resultdevice or host pointer for storing contiguous batchCount results. return is 0 if n <= 0, incx<=0.

◆ hipblasIzamin()

hipblasStatus_t hipblasIzamin ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
int *  result 
)

BLAS Level 1 API.

amin finds the first index of the element of minimum magnitude of a vector x.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the amin index. return is 0.0 if n, incx<=0.

◆ hipblasIzaminBatched()

hipblasStatus_t hipblasIzaminBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

aminBatched finds the first index of the element of minimum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch, must be > 0.
[out]resultdevice or host pointers to array of batchCount size for results. return is 0 if n, incx<=0.

◆ hipblasIzaminStridedBatched()

hipblasStatus_t hipblasIzaminStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount,
int *  result 
)

BLAS Level 1 API.

aminStridedBatched finds the first index of the element of minimum magnitude of each vector x_i in a batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] specifies the pointer increment between one x_i and the next x_(i + 1)
[in]batchCount[int] number of instances in the batch
[out]resultdevice or host pointer to array for storing contiguous batchCount results. return is 0 if n <= 0, incx<=0.

◆ hipblasNrm2BatchedEx()

hipblasStatus_t hipblasNrm2BatchedEx ( hipblasHandle_t  handle,
int  n,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
int  batchCount,
void *  result,
hipblasDatatype_t  resultType,
hipblasDatatype_t  executionType 
)

BLAS_EX API.

nrm2BatchedEx computes the euclidean norm over a batch of real or complex vectors

      result := sqrt( x_i'*x_i ) for real vectors x, for i = 1, ..., batchCount
      result := sqrt( x_i**H*x_i ) for complex vectors x, for i = 1, ..., batchCount
  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array of batchCount size for nrm2 results. return is 0.0 for each element if n <= 0, incx<=0.
[in]resultType[hipblasDatatype_t] specifies the datatype of the result.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasNrm2Ex()

hipblasStatus_t hipblasNrm2Ex ( hipblasHandle_t  handle,
int  n,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
void *  result,
hipblasDatatype_t  resultType,
hipblasDatatype_t  executionType 
)

BLAS_EX API.

nrm2Ex computes the euclidean norm of a real or complex vector

      result := sqrt( x'*x ) for real vectors
      result := sqrt( x**H*x ) for complex vectors
  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]xType[hipblasDatatype_t] specifies the datatype of the vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the nrm2 product. return is 0.0 if n, incx<=0.
[in]resultType[hipblasDatatype_t] specifies the datatype of the result.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasNrm2StridedBatchedEx()

hipblasStatus_t hipblasNrm2StridedBatchedEx ( hipblasHandle_t  handle,
int  n,
const void *  x,
hipblasDatatype_t  xType,
int  incx,
hipblasStride  stridex,
int  batchCount,
void *  result,
hipblasDatatype_t  resultType,
hipblasDatatype_t  executionType 
)

BLAS_EX API.

nrm2StridedBatchedEx computes the euclidean norm over a batch of real or complex vectors

      := sqrt( x_i'*x_i ) for real vectors x, for i = 1, ..., batchCount
      := sqrt( x_i**H*x_i ) for complex vectors, for i = 1, ..., batchCount
  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i.
[in]xdevice pointer to the first vector x_1.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array for storing contiguous batchCount results. return is 0.0 for each element if n <= 0, incx<=0.
[in]resultType[hipblasDatatype_t] specifies the datatype of the result.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasRotBatchedEx()

hipblasStatus_t hipblasRotBatchedEx ( hipblasHandle_t  handle,
int  n,
void *  x,
hipblasDatatype_t  xType,
int  incx,
void *  y,
hipblasDatatype_t  yType,
int  incy,
const void *  c,
const void *  s,
hipblasDatatype_t  csType,
int  batchCount,
hipblasDatatype_t  executionType 
)

BLAS EX API.

rotBatchedEx applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

In the case where cs_type is real: x := c * x + s * y y := c * y - s * x

In the case where cs_type is complex, the imaginary part of c is ignored: x := real(c) * x + s * y y := real(c) * y - conj(s) * x

  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice array of deivce pointers storing each vector x_i.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment between elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]yType[hipblasDatatype_t] specifies the datatype of each vector y_i.
[in]incy[int] specifies the increment between elements of each y_i.
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]csType[hipblasDatatype_t] specifies the datatype of c and s.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasRotEx()

hipblasStatus_t hipblasRotEx ( hipblasHandle_t  handle,
int  n,
void *  x,
hipblasDatatype_t  xType,
int  incx,
void *  y,
hipblasDatatype_t  yType,
int  incy,
const void *  c,
const void *  s,
hipblasDatatype_t  csType,
hipblasDatatype_t  executionType 
)

BLAS EX API.

rotEx applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to vectors x and y. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

In the case where cs_type is real: x := c * x + s * y y := c * y - s * x

In the case where cs_type is complex, the imaginary part of c is ignored: x := real(c) * x + s * y y := real(c) * y - conj(s) * x

  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice pointer storing vector x.
[in]xType[hipblasDatatype_t] specifies the datatype of vector x.
[in]incx[int] specifies the increment between elements of x.
[in,out]ydevice pointer storing vector y.
[in]yType[hipblasDatatype_t] specifies the datatype of vector y.
[in]incy[int] specifies the increment between elements of y.
[in]cdevice pointer or host pointer storing scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer storing scalar sine component of the rotation matrix.
[in]csType[hipblasDatatype_t] specifies the datatype of c and s.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasRotStridedBatchedEx()

hipblasStatus_t hipblasRotStridedBatchedEx ( hipblasHandle_t  handle,
int  n,
void *  x,
hipblasDatatype_t  xType,
int  incx,
hipblasStride  stridex,
void *  y,
hipblasDatatype_t  yType,
int  incy,
hipblasStride  stridey,
const void *  c,
const void *  s,
hipblasDatatype_t  csType,
int  batchCount,
hipblasDatatype_t  executionType 
)

BLAS Level 1 API.

rotStridedBatchedEx applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to strided batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

In the case where cs_type is real: x := c * x + s * y y := c * y - s * x

In the case where cs_type is complex, the imaginary part of c is ignored: x := real(c) * x + s * y y := real(c) * y - conj(s) * x

  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice pointer to the first vector x_1.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment between elements of each x_i.
[in]stridex[hipblasStride] specifies the increment from the beginning of x_i to the beginning of x_(i+1)
[in,out]ydevice pointer to the first vector y_1.
[in]yType[hipblasDatatype_t] specifies the datatype of each vector y_i.
[in]incy[int] specifies the increment between elements of each y_i.
[in]stridey[hipblasStride] specifies the increment from the beginning of y_i to the beginning of y_(i+1)
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]csType[hipblasDatatype_t] specifies the datatype of c and s.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasSasum()

hipblasStatus_t hipblasSasum ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
float *  result 
)

BLAS Level 1 API.

asum computes the sum of the magnitudes of elements of a real vector x, or the sum of magnitudes of the real and imaginary parts of elements if x is a complex vector.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x. incx must be > 0.
[in,out]resultdevice pointer or host pointer to store the asum product. return is 0.0 if n <= 0.

◆ hipblasSasumBatched()

hipblasStatus_t hipblasSasumBatched ( hipblasHandle_t  handle,
int  n,
const float *const  x[],
int  incx,
int  batchCount,
float *  result 
)

BLAS Level 1 API.

asumBatched computes the sum of the magnitudes of the elements in a batch of real vectors x_i, or the sum of magnitudes of the real and imaginary parts of elements if x_i is a complex vector, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch.
[out]resultdevice array or host array of batchCount size for results. return is 0.0 if n, incx<=0.

◆ hipblasSasumStridedBatched()

hipblasStatus_t hipblasSasumStridedBatched ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
hipblasStride  stridex,
int  batchCount,
float *  result 
)

BLAS Level 1 API.

asumStridedBatched computes the sum of the magnitudes of elements of a real vectors x_i, or the sum of magnitudes of the real and imaginary parts of elements if x_i is a complex vector, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array for storing contiguous batchCount results. return is 0.0 if n, incx<=0.

◆ hipblasSaxpy()

hipblasStatus_t hipblasSaxpy ( hipblasHandle_t  handle,
int  n,
const float *  alpha,
const float *  x,
int  incx,
float *  y,
int  incy 
)

BLAS Level 1 API.

axpy computes constant alpha multiplied by vector x, plus vector y

y := alpha * x + y
  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]alphadevice pointer or host pointer to specify the scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[out]ydevice pointer storing vector y.
[in,out]incy[int] specifies the increment for the elements of y.

◆ hipblasSaxpyBatched()

hipblasStatus_t hipblasSaxpyBatched ( hipblasHandle_t  handle,
int  n,
const float *  alpha,
const float *const  x[],
int  incx,
float *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

axpyBatched compute y := alpha * x + y over a set of batched vectors.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]alphaspecifies the scalar alpha.
[in]xpointer storing vector x on the GPU.
[in]incx[int] specifies the increment for the elements of x.
[out]ypointer storing vector y on the GPU.
[in,out]incy[int] specifies the increment for the elements of y.
[in]batchCount[int] number of instances in the batch

◆ hipblasSaxpyStridedBatched()

hipblasStatus_t hipblasSaxpyStridedBatched ( hipblasHandle_t  handle,
int  n,
const float *  alpha,
const float *  x,
int  incx,
hipblasStride  stridex,
float *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

axpyStridedBatched compute y := alpha * x + y over a set of strided batched vectors.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int]
[in]alphaspecifies the scalar alpha.
[in]xpointer storing vector x on the GPU.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] specifies the increment between vectors of x.
[out]ypointer storing vector y on the GPU.
[in,out]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] specifies the increment between vectors of y.
[in]batchCount[int] number of instances in the batch

◆ hipblasScalBatchedEx()

hipblasStatus_t hipblasScalBatchedEx ( hipblasHandle_t  handle,
int  n,
const void *  alpha,
hipblasDatatype_t  alphaType,
void *  x,
hipblasDatatype_t  xType,
int  incx,
int  batchCount,
hipblasDatatype_t  executionType 
)

BLAS EX API.

scalBatchedEx scales each element of each vector x_i with scalar alpha.

x_i := alpha * x_i
  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]alphadevice pointer or host pointer for the scalar alpha.
[in]alphaType[hipblasDatatype_t] specifies the datatype of alpha.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasScalEx()

hipblasStatus_t hipblasScalEx ( hipblasHandle_t  handle,
int  n,
const void *  alpha,
hipblasDatatype_t  alphaType,
void *  x,
hipblasDatatype_t  xType,
int  incx,
hipblasDatatype_t  executionType 
)

BLAS EX API.

scalEx scales each element of vector x with scalar alpha.

x := alpha * x
  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]alphadevice pointer or host pointer for the scalar alpha.
[in]alphaType[hipblasDatatype_t] specifies the datatype of alpha.
[in,out]xdevice pointer storing vector x.
[in]xType[hipblasDatatype_t] specifies the datatype of vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasScalStridedBatchedEx()

hipblasStatus_t hipblasScalStridedBatchedEx ( hipblasHandle_t  handle,
int  n,
const void *  alpha,
hipblasDatatype_t  alphaType,
void *  x,
hipblasDatatype_t  xType,
int  incx,
hipblasStride  stridex,
int  batchCount,
hipblasDatatype_t  executionType 
)

BLAS EX API.

scalStridedBatchedEx scales each element of vector x with scalar alpha over a set of strided batched vectors.

x := alpha * x

  • Supported types are determined by the backend. See rocBLAS/cuBLAS documentation.
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]alphadevice pointer or host pointer for the scalar alpha.
[in]alphaType[hipblasDatatype_t] specifies the datatype of alpha.
[in,out]xdevice pointer to the first vector x_1.
[in]xType[hipblasDatatype_t] specifies the datatype of each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) to the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size, for a typical case this means stridex >= n * incx.
[in]batchCount[int] number of instances in the batch.
[in]executionType[hipblasDatatype_t] specifies the datatype of computation.

◆ hipblasScasum()

hipblasStatus_t hipblasScasum ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
float *  result 
)

BLAS Level 1 API.

asum computes the sum of the magnitudes of elements of a real vector x, or the sum of magnitudes of the real and imaginary parts of elements if x is a complex vector.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x. incx must be > 0.
[in,out]resultdevice pointer or host pointer to store the asum product. return is 0.0 if n <= 0.

◆ hipblasScasumBatched()

hipblasStatus_t hipblasScasumBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex *const  x[],
int  incx,
int  batchCount,
float *  result 
)

BLAS Level 1 API.

asumBatched computes the sum of the magnitudes of the elements in a batch of real vectors x_i, or the sum of magnitudes of the real and imaginary parts of elements if x_i is a complex vector, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch.
[out]resultdevice array or host array of batchCount size for results. return is 0.0 if n, incx<=0.

◆ hipblasScasumStridedBatched()

hipblasStatus_t hipblasScasumStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount,
float *  result 
)

BLAS Level 1 API.

asumStridedBatched computes the sum of the magnitudes of elements of a real vectors x_i, or the sum of magnitudes of the real and imaginary parts of elements if x_i is a complex vector, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each vector x_i
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array for storing contiguous batchCount results. return is 0.0 if n, incx<=0.

◆ hipblasScnrm2()

hipblasStatus_t hipblasScnrm2 ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
float *  result 
)

BLAS Level 1 API.

nrm2 computes the euclidean norm of a real or complex vector

      result := sqrt( x'*x ) for real vectors
      result := sqrt( x**H*x ) for complex vectors
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : s,d,sc,dz
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the nrm2 product. return is 0.0 if n, incx<=0.

◆ hipblasScnrm2Batched()

hipblasStatus_t hipblasScnrm2Batched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex *const  x[],
int  incx,
int  batchCount,
float *  result 
)

BLAS Level 1 API.

nrm2Batched computes the euclidean norm over a batch of real or complex vectors

      result := sqrt( x_i'*x_i ) for real vectors x, for i = 1, ..., batchCount
      result := sqrt( x_i**H*x_i ) for complex vectors x, for i = 1, ..., batchCount
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array of batchCount size for nrm2 results. return is 0.0 for each element if n <= 0, incx<=0.

◆ hipblasScnrm2StridedBatched()

hipblasStatus_t hipblasScnrm2StridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount,
float *  result 
)

BLAS Level 1 API.

nrm2StridedBatched computes the euclidean norm over a batch of real or complex vectors

      := sqrt( x_i'*x_i ) for real vectors x, for i = 1, ..., batchCount
      := sqrt( x_i**H*x_i ) for complex vectors, for i = 1, ..., batchCount
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array for storing contiguous batchCount results. return is 0.0 for each element if n <= 0, incx<=0.

◆ hipblasScopy()

hipblasStatus_t hipblasScopy ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
float *  y,
int  incy 
)

BLAS Level 1 API.

copy copies each element x[i] into y[i], for i = 1 , ... , n

y := x,
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x to be copied to y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasScopyBatched()

hipblasStatus_t hipblasScopyBatched ( hipblasHandle_t  handle,
int  n,
const float *const  x[],
int  incx,
float *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

copyBatched copies each element x_i[j] into y_i[j], for j = 1 , ... , n; i = 1 , ... , batchCount

y_i := x_i,

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i to be copied to y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasScopyStridedBatched()

hipblasStatus_t hipblasScopyStridedBatched ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
hipblasStride  stridex,
float *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

copyStridedBatched copies each element x_i[j] into y_i[j], for j = 1 , ... , n; i = 1 , ... , batchCount

y_i := x_i,

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i to be copied to y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increments for the elements of vectors x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of vectors y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stride_y, however the user should take care to ensure that stride_y is of appropriate size, for a typical case this means stride_y >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasSdgmm()

hipblasStatus_t hipblasSdgmm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const float *  AP,
int  lda,
const float *  x,
int  incx,
float *  CP,
int  ldc 
)

BLAS Level 3 API.

dgmm performs one of the matrix-matrix operations

C = A * diag(x) if side == HIPBLAS_SIDE_RIGHT
C = diag(x) * A if side == HIPBLAS_SIDE_LEFT

where C and A are m by n dimensional matrices. diag( x ) is a diagonal matrix and x is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between values of x
[in,out]CPdevice pointer storing matrix C.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasSdgmmBatched()

hipblasStatus_t hipblasSdgmmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const float *const  AP[],
int  lda,
const float *const  x[],
int  incx,
float *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

dgmmBatched performs one of the batched matrix-matrix operations

C_i = A_i * diag(x_i) for i = 0, 1, ... batchCount-1 if side == HIPBLAS_SIDE_RIGHT
C_i = diag(x_i) * A_i for i = 0, 1, ... batchCount-1 if side == HIPBLAS_SIDE_LEFT

where C_i and A_i are m by n dimensional matrices. diag(x_i) is a diagonal matrix and x_i is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, n )
[in]lda[int] specifies the leading dimension of A_i.
[in]xdevice array of device pointers storing each vector x_i on the GPU. Each x_i is of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT
[in]incx[int] specifies the increment between values of x_i
[in,out]CPdevice array of device pointers storing each matrix C_i on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasSdgmmStridedBatched()

hipblasStatus_t hipblasSdgmmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const float *  AP,
int  lda,
hipblasStride  strideA,
const float *  x,
int  incx,
hipblasStride  stridex,
float *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

dgmmStridedBatched performs one of the batched matrix-matrix operations

C_i = A_i * diag(x_i)   if side == HIPBLAS_SIDE_RIGHT   for i = 0, 1, ... batchCount-1
C_i = diag(x_i) * A_i   if side == HIPBLAS_SIDE_LEFT    for i = 0, 1, ... batchCount-1

where C_i and A_i are m by n dimensional matrices. diag(x_i) is a diagonal matrix and x_i is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, n )
[in]lda[int] specifies the leading dimension of A.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xpointer to the first vector x_0 on the GPU. Each x_i is of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT
[in]incx[int] specifies the increment between values of x
[in]stridex[hipblasStride] stride from the start of one vector(x_i) and the next one (x_i+1)
[in,out]CPdevice pointer to the first matrix C_0 on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasSdot()

hipblasStatus_t hipblasSdot ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
const float *  y,
int  incy,
float *  result 
)

BLAS Level 1 API.

dot(u) performs the dot product of vectors x and y

result = x * y;

dotc performs the dot product of the conjugate of complex vector x and complex vector y

result = conjugate (x) * y;
  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the dot product. return is 0.0 if n <= 0.

◆ hipblasSdotBatched()

hipblasStatus_t hipblasSdotBatched ( hipblasHandle_t  handle,
int  n,
const float *const  x[],
int  incx,
const float *const  y[],
int  incy,
int  batchCount,
float *  result 
)

BLAS Level 1 API.

dotBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasSdotStridedBatched()

hipblasStatus_t hipblasSdotStridedBatched ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
hipblasStride  stridex,
const float *  y,
int  incy,
hipblasStride  stridey,
int  batchCount,
float *  result 
)

BLAS Level 1 API.

dotStridedBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcStridedBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasSetAtomicsMode()

hipblasStatus_t hipblasSetAtomicsMode ( hipblasHandle_t  handle,
hipblasAtomicsMode_t  atomics_mode 
)

Set hipblasSetAtomicsMode.

◆ hipblasSetMatrix()

hipblasStatus_t hipblasSetMatrix ( int  rows,
int  cols,
int  elemSize,
const void *  AP,
int  lda,
void *  BP,
int  ldb 
)

copy matrix from host to device

Parameters
[in]rows[int] number of rows in matrices
[in]cols[int] number of columns in matrices
[in]elemSize[int] number of bytes per element in the matrix
[in]APpointer to matrix on the host
[in]lda[int] specifies the leading dimension of A, lda >= rows
[out]BPpointer to matrix on the GPU
[in]ldb[int] specifies the leading dimension of B, ldb >= rows

◆ hipblasSetMatrixAsync()

hipblasStatus_t hipblasSetMatrixAsync ( int  rows,
int  cols,
int  elemSize,
const void *  AP,
int  lda,
void *  BP,
int  ldb,
hipStream_t  stream 
)

asynchronously copy matrix from host to device

hipblasSetMatrixAsync copies a matrix from pinned host memory to device memory asynchronously. Memory on the host must be allocated with hipHostMalloc or the transfer will be synchronous.

Parameters
[in]rows[int] number of rows in matrices
[in]cols[int] number of columns in matrices
[in]elemSize[int] number of bytes per element in the matrix
[in]APpointer to matrix on the host
[in]lda[int] specifies the leading dimension of A, lda >= rows
[out]BPpointer to matrix on the GPU
[in]ldb[int] specifies the leading dimension of B, ldb >= rows
[in]streamspecifies the stream into which this transfer request is queued

◆ hipblasSetPointerMode()

hipblasStatus_t hipblasSetPointerMode ( hipblasHandle_t  handle,
hipblasPointerMode_t  mode 
)

Set hipblas pointer mode.

◆ hipblasSetStream()

hipblasStatus_t hipblasSetStream ( hipblasHandle_t  handle,
hipStream_t  streamId 
)

Set stream for handle.

◆ hipblasSetVector()

hipblasStatus_t hipblasSetVector ( int  n,
int  elemSize,
const void *  x,
int  incx,
void *  y,
int  incy 
)

copy vector from host to device

Parameters
[in]n[int] number of elements in the vector
[in]elemSize[int] Size of both vectors in bytes
[in]xpointer to vector on the host
[in]incx[int] specifies the increment for the elements of the vector
[out]ypointer to vector on the device
[in]incy[int] specifies the increment for the elements of the vector

◆ hipblasSetVectorAsync()

hipblasStatus_t hipblasSetVectorAsync ( int  n,
int  elemSize,
const void *  x,
int  incx,
void *  y,
int  incy,
hipStream_t  stream 
)

asynchronously copy vector from host to device

hipblasSetVectorAsync copies a vector from pinned host memory to device memory asynchronously. Memory on the host must be allocated with hipHostMalloc or the transfer will be synchronous.

Parameters
[in]n[int] number of elements in the vector
[in]elemSize[int] number of bytes per element in the matrix
[in]xpointer to vector on the host
[in]incx[int] specifies the increment for the elements of the vector
[out]ypointer to vector on the device
[in]incy[int] specifies the increment for the elements of the vector
[in]streamspecifies the stream into which this transfer request is queued

◆ hipblasSgbmv()

hipblasStatus_t hipblasSgbmv ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const float *  alpha,
const float *  AP,
int  lda,
const float *  x,
int  incx,
const float *  beta,
float *  y,
int  incy 
)

BLAS Level 2 API.

gbmv performs one of the matrix-vector operations

y := alpha*A*x    + beta*y,   or
y := alpha*A**T*x + beta*y,   or
y := alpha*A**H*x + beta*y,

where alpha and beta are scalars, x and y are vectors and A is an m by n banded matrix with kl sub-diagonals and ku super-diagonals.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]kl[int] number of sub-diagonals of A
[in]ku[int] number of super-diagonals of A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing banded matrix A. Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of A. Must be >= (kl + ku + 1)
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasSgbmvBatched()

hipblasStatus_t hipblasSgbmvBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const float *  alpha,
const float *const  AP[],
int  lda,
const float *const  x[],
int  incx,
const float *  beta,
float *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

gbmvBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n banded matrix with kl sub-diagonals and ku super-diagonals, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of each matrix A_i
[in]n[int] number of columns of each matrix A_i
[in]kl[int] number of sub-diagonals of each A_i
[in]ku[int] number of super-diagonals of each A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each banded matrix A_i. Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of each A_i. Must be >= (kl + ku + 1)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasSgbmvStridedBatched()

hipblasStatus_t hipblasSgbmvStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
const float *  x,
int  incx,
hipblasStride  stridex,
const float *  beta,
float *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

gbmvStridedBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n banded matrix with kl sub-diagonals and ku super-diagonals, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]kl[int] number of sub-diagonals of A
[in]ku[int] number of super-diagonals of A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer to first banded matrix (A_1). Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of A. Must be >= (kl + ku + 1)
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xdevice pointer to first vector (x_1).
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer to first vector (y_1).
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (x_i+1)
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasSgeam()

hipblasStatus_t hipblasSgeam ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
const float *  beta,
const float *  BP,
int  ldb,
float *  CP,
int  ldc 
)

BLAS Level 3 API.

geam performs one of the matrix-matrix operations

C = alpha*op( A ) + beta*op( B ),

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are matrices, with op( A ) an m by n matrix, op( B ) an m by n matrix, and C an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPdevice pointer storing matrix B.
[in]ldb[int] specifies the leading dimension of B.
[in,out]CPdevice pointer storing matrix C.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasSgeamBatched()

hipblasStatus_t hipblasSgeamBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const float *  alpha,
const float *const  AP[],
int  lda,
const float *  beta,
const float *const  BP[],
int  ldb,
float *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

geamBatched performs one of the batched matrix-matrix operations

C_i = alpha*op( A_i ) + beta*op( B_i )  for i = 0, 1, ... batchCount - 1

where alpha and beta are scalars, and op(A_i), op(B_i) and C_i are m by n matrices and op( X ) is one of

op( X ) = X      or
op( X ) = X**T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, k ), where k is m when transA == HIPBLAS_OP_N and is n when transA == HIPBLAS_OP_T.
[in]lda[int] specifies the leading dimension of A.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPdevice array of device pointers storing each matrix B_i on the GPU. Each B_i is of dimension ( ldb, k ), where k is m when transB == HIPBLAS_OP_N and is n when transB == HIPBLAS_OP_T.
[in]ldb[int] specifies the leading dimension of B.
[in,out]CPdevice array of device pointers storing each matrix C_i on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]batchCount[int] number of instances i in the batch.

◆ hipblasSgeamStridedBatched()

hipblasStatus_t hipblasSgeamStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
const float *  beta,
const float *  BP,
int  ldb,
hipblasStride  strideB,
float *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

geamStridedBatched performs one of the batched matrix-matrix operations

C_i = alpha*op( A_i ) + beta*op( B_i )  for i = 0, 1, ... batchCount - 1

where alpha and beta are scalars, and op(A_i), op(B_i) and C_i are m by n matrices and op( X ) is one of

op( X ) = X      or
op( X ) = X**T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, k ), where k is m when transA == HIPBLAS_OP_N and is n when transA == HIPBLAS_OP_T.
[in]lda[int] specifies the leading dimension of A.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPpointer to the first matrix B_0 on the GPU. Each B_i is of dimension ( ldb, k ), where k is m when transB == HIPBLAS_OP_N and is n when transB == HIPBLAS_OP_T.
[in]ldb[int] specifies the leading dimension of B.
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in,out]CPpointer to the first matrix C_0 on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasSgels()

hipblasStatus_t hipblasSgels ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
float *  A,
const int  lda,
float *  B,
const int  ldb,
int *  info,
int *  deviceInfo 
)

GELS solves an overdetermined (or underdetermined) linear system defined by an m-by-n matrix A, and a corresponding matrix B, using the QR factorization computed by GEQRF (or the LQ factorization computed by "GELQF").

Depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A X = B & \: \text{not transposed, or}\\ A' X = B & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X is found by minimizing

\[ || B - A X || \quad \text{(or} \: || B - A' X ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X is chosen such that \(|| X ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : currently unsupported
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of matrix A.
[in]nint. n >= 0.
The number of columns of matrix A.
[in]nrhsint. nrhs >= 0.
The number of columns of matrices B and X; i.e., the columns on the right hand side.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the matrix A. On exit, the QR (or LQ) factorization of A as returned by "GEQRF" (or "GELQF").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrix A.
[in,out]Bpointer to type. Array on the GPU of dimension ldb*nrhs.
On entry, the matrix B. On exit, when info = 0, B is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrix B.
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int on the GPU.
If info = 0, successful exit. If info = i > 0, the solution could not be computed because input matrix A is rank deficient; the i-th diagonal element of its triangular factor is zero.

◆ hipblasSgelsBatched()

hipblasStatus_t hipblasSgelsBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
float *const  A[],
const int  lda,
float *const  B[],
const int  ldb,
int *  info,
int *  deviceInfo,
const int  batchCount 
)

gelsBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_BATCHED" (or the LQ factorizations computed by "GELQF_BATCHED").

For each instance in the batch, depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A_j X_j = B_j & \: \text{not transposed, or}\\ A_j' X_j = B_j & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X_j is found by minimizing

\[ || B_j - A_j X_j || \quad \text{(or} \: || B_j - A_j' X_j ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X_j is chosen such that \(|| X_j ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z Note that cuBLAS backend supports only the non-transpose operation and only solves over-determined systems (m >= n).
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of all matrices A_j in the batch.
[in]nint. n >= 0.
The number of columns of all matrices A_j in the batch.
[in]nrhsint. nrhs >= 0.
The number of columns of all matrices B_j and X_j in the batch; i.e., the columns on the right hand side.
[in,out]Aarray of pointer to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the matrices A_j. On exit, the QR (or LQ) factorizations of A_j as returned by "GEQRF_BATCHED" (or "GELQF_BATCHED").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_j.
[in,out]Barray of pointer to type. Each pointer points to an array on the GPU of dimension ldb*nrhs.
On entry, the matrices B_j. On exit, when info[j] = 0, B_j is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrices B_j.
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int. Array of batchCount integers on the GPU.
If deviceInfo[j] = 0, successful exit for solution of A_j. If deviceInfo[j] = i > 0, the solution of A_j could not be computed because input matrix A_j is rank deficient; the i-th diagonal element of its triangular factor is zero.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasSgelsStridedBatched()

hipblasStatus_t hipblasSgelsStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
float *  A,
const int  lda,
const hipblasStride  strideA,
float *  B,
const int  ldb,
const hipblasStride  strideB,
int *  info,
int *  deviceInfo,
const int  batch_count 
)

gelsStridedBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_STRIDED_BATCHED" (or the LQ factorizations computed by "GELQF_STRIDED_BATCHED").

For each instance in the batch, depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A_j X_j = B_j & \: \text{not transposed, or}\\ A_j' X_j = B_j & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X_j is found by minimizing

\[ || B_j - A_j X_j || \quad \text{(or} \: || B_j - A_j' X_j ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X_j is chosen such that \(|| X_j ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : currently unsupported
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of all matrices A_j in the batch.
[in]nint. n >= 0.
The number of columns of all matrices A_j in the batch.
[in]nrhsint. nrhs >= 0.
The number of columns of all matrices B_j and X_j in the batch; i.e., the columns on the right hand side.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the matrices A_j. On exit, the QR (or LQ) factorizations of A_j as returned by "GEQRF_STRIDED_BATCHED" (or "GELQF_STRIDED_BATCHED").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_j.
[in]strideAhipblasStride.
Stride from the start of one matrix A_j to the next one A_(j+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n
[in,out]Bpointer to type. Array on the GPU (the size depends on the value of strideB).
On entry, the matrices B_j. On exit, when info[j] = 0, each B_j is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrices B_j.
[in]strideBhipblasStride.
Stride from the start of one matrix B_j to the next one B_(j+1). There is no restriction for the value of strideB. Normal use case is strideB >= ldb*nrhs
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int. Array of batchCount integers on the GPU.
If deviceInfo[j] = 0, successful exit for solution of A_j. If deviceInfo[j] = i > 0, the solution of A_j could not be computed because input matrix A_j is rank deficient; the i-th diagonal element of its triangular factor is zero.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasSgemm()

hipblasStatus_t hipblasSgemm ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const float *  alpha,
const float *  AP,
int  lda,
const float *  BP,
int  ldb,
const float *  beta,
float *  CP,
int  ldc 
)

BLAS Level 3 API.

gemm performs one of the matrix-matrix operations

C = alpha*op( A )*op( B ) + beta*C,

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are matrices, with op( A ) an m by k matrix, op( B ) a k by n matrix and C an m by n matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t]
      .
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] number or rows of matrices op( A ) and C
[in]n[int] number of columns of matrices op( B ) and C
[in]k[int] number of columns of matrix op( A ) and number of rows of matrix op( B )
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]BPdevice pointer storing matrix B.
[in]ldb[int] specifies the leading dimension of B.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice pointer storing matrix C on the GPU.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasSgemmBatched()

hipblasStatus_t hipblasSgemmBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const float *  alpha,
const float *const  AP[],
int  lda,
const float *const  BP[],
int  ldb,
const float *  beta,
float *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

gemmBatched performs one of the batched matrix-matrix operations C_i = alpha*op( A_i )*op( B_i ) + beta*C_i, for i = 1, ..., batchCount. where op( X ) is one of op( X ) = X or op( X ) = X**T or op( X ) = X**H, alpha and beta are scalars, and A, B and C are strided batched matrices, with op( A ) an m by k by batchCount strided_batched matrix, op( B ) an k by n by batchCount strided_batched matrix and C an m by n by batchCount strided_batched matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimention m.
[in]n[int] matrix dimention n.
[in]k[int] matrix dimention k.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]BPdevice array of device pointers storing each matrix B_i.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice array of device pointers storing each matrix C_i.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]batchCount[int] number of gemm operations in the batch

◆ hipblasSgemmStridedBatched()

hipblasStatus_t hipblasSgemmStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const float *  alpha,
const float *  AP,
int  lda,
long long  strideA,
const float *  BP,
int  ldb,
long long  strideB,
const float *  beta,
float *  CP,
int  ldc,
long long  strideC,
int  batchCount 
)

BLAS Level 3 API.

gemmStridedBatched performs one of the strided batched matrix-matrix operations

C_i = alpha*op( A_i )*op( B_i ) + beta*C_i, for i = 1, ..., batchCount.

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are strided batched matrices, with op( A ) an m by k by batchCount strided_batched matrix, op( B ) an k by n by batchCount strided_batched matrix and C an m by n by batchCount strided_batched matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimention m.
[in]n[int] matrix dimention n.
[in]k[int] matrix dimention k.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer pointing to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in]BPdevice pointer pointing to the first matrix B_1.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]strideB[hipblasStride] stride from the start of one B_i matrix to the next B_(i + 1).
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice pointer pointing to the first matrix C_1.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]strideC[hipblasStride] stride from the start of one C_i matrix to the next C_(i + 1).
[in]batchCount[int] number of gemm operatons in the batch

◆ hipblasSgemv()

hipblasStatus_t hipblasSgemv ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
const float *  x,
int  incx,
const float *  beta,
float *  y,
int  incy 
)

BLAS Level 2 API.

gemv performs one of the matrix-vector operations

y := alpha*A*x    + beta*y,   or
y := alpha*A**T*x + beta*y,   or
y := alpha*A**H*x + beta*y,

where alpha and beta are scalars, x and y are vectors and A is an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasSgemvBatched()

hipblasStatus_t hipblasSgemvBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
const float *  alpha,
const float *const  AP[],
int  lda,
const float *const  x[],
int  incx,
const float *  beta,
float *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

gemvBatched performs a batch of matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrices A_i are tranposed (conjugated) or not
[in]m[int] number of rows of each matrix A_i
[in]n[int] number of columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each matrix A_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasSgemvStridedBatched()

hipblasStatus_t hipblasSgemvStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
const float *  x,
int  incx,
hipblasStride  stridex,
const float *  beta,
float *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

gemvStridedBatched performs a batch of matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] indicates whether matrices A_i are tranposed (conjugated) or not
[in]m[int] number of rows of matrices A_i
[in]n[int] number of columns of matrices A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer to the first matrix (A_1) in the batch.
[in]lda[int] specifies the leading dimension of matrices A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. When trans equals HIPBLAS_OP_N this typically means stridex >= n * incx, otherwise stridex >= m * incx.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of vectors y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. When trans equals HIPBLAS_OP_N this typically means stridey >= m * incy, otherwise stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasSgeqrf()

hipblasStatus_t hipblasSgeqrf ( hipblasHandle_t  handle,
const int  m,
const int  n,
float *  A,
const int  lda,
float *  ipiv,
int *  info 
)

SOLVER API.

geqrf computes a QR factorization of a general m-by-n matrix A.

The factorization has the form

\[ A = Q\left[\begin{array}{c} R\\ 0 \end{array}\right] \]

where R is upper triangular (upper trapezoidal if m < n), and Q is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q = H_1H_2\cdots H_k, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_i\) is given by

\[ H_i = I - \text{ipiv}[i] \cdot v_i v_i' \]

where the first i-1 elements of the Householder vector \(v_i\) are zero, and \(v_i[i] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of the matrix A.
[in]nint. n >= 0.
The number of columns of the matrix A.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the m-by-n matrix to be factored. On exit, the elements on and above the diagonal contain the factor R; the elements below the diagonal are the last m - i elements of Householder vector v_i.
[in]ldaint. lda >= m.
Specifies the leading dimension of A.
[out]ipivpointer to type. Array on the GPU of dimension min(m,n).
The Householder scalars.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.

◆ hipblasSgeqrfBatched()

hipblasStatus_t hipblasSgeqrfBatched ( hipblasHandle_t  handle,
const int  m,
const int  n,
float *const  A[],
const int  lda,
float *const  ipiv[],
int *  info,
const int  batchCount 
)

SOLVER API.

geqrfBatched computes the QR factorization of a batch of general m-by-n matrices.

The factorization of matrix \(A_i\) in the batch has the form

\[ A_i = Q_i\left[\begin{array}{c} R_i\\ 0 \end{array}\right] \]

where \(R_i\) is upper triangular (upper trapezoidal if m < n), and \(Q_i\) is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q_i = H_{i_1}H_{i_2}\cdots H_{i_k}, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_{i_j}\) is given by

\[ H_{i_j} = I - \text{ipiv}_i[j] \cdot v_{i_j} v_{i_j}' \]

where the first j-1 elements of Householder vector \(v_{i_j}\) are zero, and \(v_{i_j}[j] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of all the matrices A_i in the batch.
[in]nint. n >= 0.
The number of columns of all the matrices A_i in the batch.
[in,out]AArray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the m-by-n matrices A_i to be factored. On exit, the elements on and above the diagonal contain the factor R_i. The elements below the diagonal are the last m - j elements of Householder vector v_(i_j).
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_i.
[out]ipivarray of pointers to type. Each pointer points to an array on the GPU of dimension min(m, n).
Contains the vectors ipiv_i of corresponding Householder scalars.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasSgeqrfStridedBatched()

hipblasStatus_t hipblasSgeqrfStridedBatched ( hipblasHandle_t  handle,
const int  m,
const int  n,
float *  A,
const int  lda,
const hipblasStride  strideA,
float *  ipiv,
const hipblasStride  strideP,
int *  info,
const int  batchCount 
)

SOLVER API.

geqrfStridedBatched computes the QR factorization of a batch of general m-by-n matrices.

The factorization of matrix \(A_i\) in the batch has the form

\[ A_i = Q_i\left[\begin{array}{c} R_i\\ 0 \end{array}\right] \]

where \(R_i\) is upper triangular (upper trapezoidal if m < n), and \(Q_i\) is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q_i = H_{i_1}H_{i_2}\cdots H_{i_k}, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_{i_j}\) is given by

\[ H_{i_j} = I - \text{ipiv}_j[j] \cdot v_{i_j} v_{i_j}' \]

where the first j-1 elements of Householder vector \(v_{i_j}\) are zero, and \(v_{i_j}[j] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of all the matrices A_i in the batch.
[in]nint. n >= 0.
The number of columns of all the matrices A_i in the batch.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the m-by-n matrices A_i to be factored. On exit, the elements on and above the diagonal contain the factor R_i. The elements below the diagonal are the last m - j elements of Householder vector v_(i_j).
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n.
[out]ipivpointer to type. Array on the GPU (the size depends on the value of strideP).
Contains the vectors ipiv_i of corresponding Householder scalars.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use is strideP >= min(m,n).
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasSger()

hipblasStatus_t hipblasSger ( hipblasHandle_t  handle,
int  m,
int  n,
const float *  alpha,
const float *  x,
int  incx,
const float *  y,
int  incy,
float *  AP,
int  lda 
)

BLAS Level 2 API.

ger,geru,gerc performs the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where alpha is a scalar, x and y are vectors, and A is an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of the matrix A.
[in]n[int] the number of columns of the matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasSgerBatched()

hipblasStatus_t hipblasSgerBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const float *  alpha,
const float *const  x[],
int  incx,
const float *const  y[],
int  incy,
float *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

gerBatched,geruBatched,gercBatched performs a batch of the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of eaceh matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasSgerStridedBatched()

hipblasStatus_t hipblasSgerStridedBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const float *  alpha,
const float *  x,
int  incx,
hipblasStride  stridex,
const float *  y,
int  incy,
hipblasStride  stridey,
float *  AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

gerStridedBatched,geruStridedBatched,gercStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**T, OR
A_i := A_i + alpha*x_i*y_i**H  for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increments for the elements of each vector x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size, for a typical case this means stridex >= m * incx.
[in,out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size, for a typical case this means stridey >= n * incy.
[in,out]APdevice pointer to the first matrix (A_1) in the batch.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasSgetrf()

hipblasStatus_t hipblasSgetrf ( hipblasHandle_t  handle,
const int  n,
float *  A,
const int  lda,
int *  ipiv,
int *  info 
)

SOLVER API.

getrf computes the LU factorization of a general n-by-n matrix A using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization has the form:

\[ A = PLU \]

where P is a permutation matrix, L is lower triangular with unit diagonal elements, and U is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A = LU \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of the matrix A.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the n-by-n matrix A to be factored. On exit, the factors L and U from the factorization. The unit diagonal elements of L are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of A.
[out]ipivpointer to int. Array on the GPU of dimension n.
The vector of pivot indices. Elements of ipiv are 1-based indices. For 1 <= i <= n, the row i of the matrix was interchanged with row ipiv[i]. Matrix P of the factorization can be derived from ipiv. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[out]infopointer to a int on the GPU.
If info = 0, successful exit. If info = j > 0, U is singular. U[j,j] is the first zero pivot.

◆ hipblasSgetrfBatched()

hipblasStatus_t hipblasSgetrfBatched ( hipblasHandle_t  handle,
const int  n,
float *const  A[],
const int  lda,
int *  ipiv,
int *  info,
const int  batchCount 
)

SOLVER API.

getrfBatched computes the LU factorization of a batch of general n-by-n matrices using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization of matrix \(A_i\) in the batch has the form:

\[ A_i = P_iL_iU_i \]

where \(P_i\) is a permutation matrix, \(L_i\) is lower triangular with unit diagonal elements, and \(U_i\) is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A_i = L_iU_i \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of all matrices A_i in the batch.
[in,out]Aarray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the n-by-n matrices A_i to be factored. On exit, the factors L_i and U_i from the factorizations. The unit diagonal elements of L_i are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[out]ipivpointer to int. Array on the GPU.
Contains the vectors of pivot indices ipiv_i (corresponding to A_i). Dimension of ipiv_i is n. Elements of ipiv_i are 1-based indices. For each instance A_i in the batch and for 1 <= j <= n, the row j of the matrix A_i was interchanged with row ipiv_i[j]. Matrix P_i of the factorization can be derived from ipiv_i. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for factorization of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasSgetrfStridedBatched()

hipblasStatus_t hipblasSgetrfStridedBatched ( hipblasHandle_t  handle,
const int  n,
float *  A,
const int  lda,
const hipblasStride  strideA,
int *  ipiv,
const hipblasStride  strideP,
int *  info,
const int  batchCount 
)

SOLVER API.

getrfStridedBatched computes the LU factorization of a batch of general n-by-n matrices using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization of matrix \(A_i\) in the batch has the form:

\[ A_i = P_iL_iU_i \]

where \(P_i\) is a permutation matrix, \(L_i\) is lower triangular with unit diagonal elements, and \(U_i\) is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A_i = L_iU_i \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of all matrices A_i in the batch.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the n-by-n matrices A_i to be factored. On exit, the factors L_i and U_i from the factorization. The unit diagonal elements of L_i are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n
[out]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
Contains the vectors of pivots indices ipiv_i (corresponding to A_i). Dimension of ipiv_i is n. Elements of ipiv_i are 1-based indices. For each instance A_i in the batch and for 1 <= j <= n, the row j of the matrix A_i was interchanged with row ipiv_i[j]. Matrix P_i of the factorization can be derived from ipiv_i. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use case is strideP >= n.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for factorization of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasSgetriBatched()

hipblasStatus_t hipblasSgetriBatched ( hipblasHandle_t  handle,
const int  n,
float *const  A[],
const int  lda,
int *  ipiv,
float *const  C[],
const int  ldc,
int *  info,
const int  batchCount 
)

SOLVER API.

getriBatched computes the inverse \(C_i = A_i^{-1}\) of a batch of general n-by-n matrices \(A_i\).

The inverse is computed by solving the linear system

\[ A_i C_i = I \]

where I is the identity matrix, and \(A_i\) is factorized as \(A_i = P_i L_i U_i\) as given by getrfBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of rows and columns of all matrices A_i in the batch.
[in]Aarray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfBatched.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[in]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
The pivot indices returned by getrfBatched. ipiv can be passed in as a nullptr, this will assume that getrfBatched was called without partial pivoting.
[out]Carray of pointers to type. Each pointer points to an array on the GPU of dimension ldc*n.
If info[i] = 0, the inverse of matrices A_i. Otherwise, undefined.
[in]ldcint. ldc >= n.
Specifies the leading dimension of C_i.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for inversion of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasSgetrs()

hipblasStatus_t hipblasSgetrs ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
float *  A,
const int  lda,
const int *  ipiv,
float *  B,
const int  ldb,
int *  info 
)

SOLVER API.

getrs solves a system of n linear equations on n variables in its factorized form.

It solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A X = B & \: \text{not transposed,}\\ A^T X = B & \: \text{transposed, or}\\ A^H X = B & \: \text{conjugate transposed.} \end{array} \]

Matrix A is defined by its triangular factors as returned by getrf.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of A.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of the matrix B.
[in]Apointer to type. Array on the GPU of dimension lda*n.
The factors L and U of the factorization A = P*L*U returned by getrf.
[in]ldaint. lda >= n.
The leading dimension of A.
[in]ipivpointer to int. Array on the GPU of dimension n.
The pivot indices returned by getrf.
[in,out]Bpointer to type. Array on the GPU of dimension ldb*nrhs.
On entry, the right hand side matrix B. On exit, the solution matrix X.
[in]ldbint. ldb >= n.
The leading dimension of B.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.

◆ hipblasSgetrsBatched()

hipblasStatus_t hipblasSgetrsBatched ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
float *const  A[],
const int  lda,
const int *  ipiv,
float *const  B[],
const int  ldb,
int *  info,
const int  batchCount 
)

SOLVER API.

getrsBatched solves a batch of systems of n linear equations on n variables in its factorized forms.

For each instance i in the batch, it solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A_i X_i = B_i & \: \text{not transposed,}\\ A_i^T X_i = B_i & \: \text{transposed, or}\\ A_i^H X_i = B_i & \: \text{conjugate transposed.} \end{array} \]

Matrix \(A_i\) is defined by its triangular factors as returned by getrfBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations of each instance in the batch.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of all A_i matrices.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of all the matrices B_i.
[in]AArray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfBatched.
[in]ldaint. lda >= n.
The leading dimension of matrices A_i.
[in]ipivpointer to int. Array on the GPU.
Contains the vectors ipiv_i of pivot indices returned by getrfBatched.
[in,out]BArray of pointers to type. Each pointer points to an array on the GPU of dimension ldb*nrhs.
On entry, the right hand side matrices B_i. On exit, the solution matrix X_i of each system in the batch.
[in]ldbint. ldb >= n.
The leading dimension of matrices B_i.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of instances (systems) in the batch.

◆ hipblasSgetrsStridedBatched()

hipblasStatus_t hipblasSgetrsStridedBatched ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
float *  A,
const int  lda,
const hipblasStride  strideA,
const int *  ipiv,
const hipblasStride  strideP,
float *  B,
const int  ldb,
const hipblasStride  strideB,
int *  info,
const int  batchCount 
)

SOLVER API.

getrsStridedBatched solves a batch of systems of n linear equations on n variables in its factorized forms.

For each instance i in the batch, it solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A_i X_i = B_i & \: \text{not transposed,}\\ A_i^T X_i = B_i & \: \text{transposed, or}\\ A_i^H X_i = B_i & \: \text{conjugate transposed.} \end{array} \]

Matrix \(A_i\) is defined by its triangular factors as returned by getrfStridedBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations of each instance in the batch.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of all A_i matrices.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of all the matrices B_i.
[in]Apointer to type. Array on the GPU (the size depends on the value of strideA).
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfStridedBatched.
[in]ldaint. lda >= n.
The leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n.
[in]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
Contains the vectors ipiv_i of pivot indices returned by getrfStridedBatched.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use case is strideP >= n.
[in,out]Bpointer to type. Array on the GPU (size depends on the value of strideB).
On entry, the right hand side matrices B_i. On exit, the solution matrix X_i of each system in the batch.
[in]ldbint. ldb >= n.
The leading dimension of matrices B_i.
[in]strideBhipblasStride.
Stride from the start of one matrix B_i to the next one B_(i+1). There is no restriction for the value of strideB. Normal use case is strideB >= ldb*nrhs.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of instances (systems) in the batch.

◆ hipblasSnrm2()

hipblasStatus_t hipblasSnrm2 ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
float *  result 
)

BLAS Level 1 API.

nrm2 computes the euclidean norm of a real or complex vector

      result := sqrt( x'*x ) for real vectors
      result := sqrt( x**H*x ) for complex vectors
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : s,d,sc,dz
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the nrm2 product. return is 0.0 if n, incx<=0.

◆ hipblasSnrm2Batched()

hipblasStatus_t hipblasSnrm2Batched ( hipblasHandle_t  handle,
int  n,
const float *const  x[],
int  incx,
int  batchCount,
float *  result 
)

BLAS Level 1 API.

nrm2Batched computes the euclidean norm over a batch of real or complex vectors

      result := sqrt( x_i'*x_i ) for real vectors x, for i = 1, ..., batchCount
      result := sqrt( x_i**H*x_i ) for complex vectors x, for i = 1, ..., batchCount
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array of batchCount size for nrm2 results. return is 0.0 for each element if n <= 0, incx<=0.

◆ hipblasSnrm2StridedBatched()

hipblasStatus_t hipblasSnrm2StridedBatched ( hipblasHandle_t  handle,
int  n,
const float *  x,
int  incx,
hipblasStride  stridex,
int  batchCount,
float *  result 
)

BLAS Level 1 API.

nrm2StridedBatched computes the euclidean norm over a batch of real or complex vectors

      := sqrt( x_i'*x_i ) for real vectors x, for i = 1, ..., batchCount
      := sqrt( x_i**H*x_i ) for complex vectors, for i = 1, ..., batchCount
  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i. incx must be > 0.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] number of instances in the batch
[out]resultdevice pointer or host pointer to array for storing contiguous batchCount results. return is 0.0 for each element if n <= 0, incx<=0.

◆ hipblasSrot()

hipblasStatus_t hipblasSrot ( hipblasHandle_t  handle,
int  n,
float *  x,
int  incx,
float *  y,
int  incy,
const float *  c,
const float *  s 
)

BLAS Level 1 API.

rot applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to vectors x and y. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment between elements of y.
[in]cdevice pointer or host pointer storing scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer storing scalar sine component of the rotation matrix.

◆ hipblasSrotBatched()

hipblasStatus_t hipblasSrotBatched ( hipblasHandle_t  handle,
int  n,
float *const  x[],
int  incx,
float *const  y[],
int  incy,
const float *  c,
const float *  s,
int  batchCount 
)

BLAS Level 1 API.

rotBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice array of deivce pointers storing each vector x_i.
[in]incx[int] specifies the increment between elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment between elements of each y_i.
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasSrotg()

hipblasStatus_t hipblasSrotg ( hipblasHandle_t  handle,
float *  a,
float *  b,
float *  c,
float *  s 
)

BLAS Level 1 API.

rotg creates the Givens rotation matrix for the vector (a b). Scalars c and s and arrays a and b may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice pointer or host pointer to input vector element, overwritten with r.
[in,out]bdevice pointer or host pointer to input vector element, overwritten with z.
[in,out]cdevice pointer or host pointer to cosine element of Givens rotation.
[in,out]sdevice pointer or host pointer sine element of Givens rotation.

◆ hipblasSrotgBatched()

hipblasStatus_t hipblasSrotgBatched ( hipblasHandle_t  handle,
float *const  a[],
float *const  b[],
float *const  c[],
float *const  s[],
int  batchCount 
)

BLAS Level 1 API.

rotgBatched creates the Givens rotation matrix for the batched vectors (a_i b_i), for i = 1, ..., batchCount. a, b, c, and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice array of device pointers storing each single input vector element a_i, overwritten with r_i.
[in,out]bdevice array of device pointers storing each single input vector element b_i, overwritten with z_i.
[in,out]cdevice array of device pointers storing each cosine element of Givens rotation for the batch.
[in,out]sdevice array of device pointers storing each sine element of Givens rotation for the batch.
[in]batchCount[int] number of batches (length of arrays a, b, c, and s).

◆ hipblasSrotgStridedBatched()

hipblasStatus_t hipblasSrotgStridedBatched ( hipblasHandle_t  handle,
float *  a,
hipblasStride  stridea,
float *  b,
hipblasStride  strideb,
float *  c,
hipblasStride  stridec,
float *  s,
hipblasStride  strides,
int  batchCount 
)

BLAS Level 1 API.

rotgStridedBatched creates the Givens rotation matrix for the strided batched vectors (a_i b_i), for i = 1, ..., batchCount. a, b, c, and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice strided_batched pointer or host strided_batched pointer to first single input vector element a_1, overwritten with r.
[in]stridea[hipblasStride] distance between elements of a in batch (distance between a_i and a_(i + 1))
[in,out]bdevice strided_batched pointer or host strided_batched pointer to first single input vector element b_1, overwritten with z.
[in]strideb[hipblasStride] distance between elements of b in batch (distance between b_i and b_(i + 1))
[in,out]cdevice strided_batched pointer or host strided_batched pointer to first cosine element of Givens rotations c_1.
[in]stridec[hipblasStride] distance between elements of c in batch (distance between c_i and c_(i + 1))
[in,out]sdevice strided_batched pointer or host strided_batched pointer to sine element of Givens rotations s_1.
[in]strides[hipblasStride] distance between elements of s in batch (distance between s_i and s_(i + 1))
[in]batchCount[int] number of batches (length of arrays a, b, c, and s).

◆ hipblasSrotm()

hipblasStatus_t hipblasSrotm ( hipblasHandle_t  handle,
int  n,
float *  x,
int  incx,
float *  y,
int  incy,
const float *  param 
)

BLAS Level 1 API.

rotm applies the modified Givens rotation matrix defined by param to vectors x and y.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : s,d
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment between elements of y.
[in]paramdevice vector or host vector of 5 elements defining the rotation. param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

◆ hipblasSrotmBatched()

hipblasStatus_t hipblasSrotmBatched ( hipblasHandle_t  handle,
int  n,
float *const  x[],
int  incx,
float *const  y[],
int  incy,
const float *const  param[],
int  batchCount 
)

BLAS Level 1 API.

rotmBatched applies the modified Givens rotation matrix defined by param_i to batched vectors x_i and y_i, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment between elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_1.
[in]incy[int] specifies the increment between elements of each y_i.
[in]paramdevice array of device vectors of 5 elements defining the rotation. param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may ONLY be stored on the device for the batched version of this function.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasSrotmg()

hipblasStatus_t hipblasSrotmg ( hipblasHandle_t  handle,
float *  d1,
float *  d2,
float *  x1,
const float *  y1,
float *  param 
)

BLAS Level 1 API.

rotmg creates the modified Givens rotation matrix for the vector (d1 * x1, d2 * y1). Parameters may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : s,d
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]d1device pointer or host pointer to input scalar that is overwritten.
[in,out]d2device pointer or host pointer to input scalar that is overwritten.
[in,out]x1device pointer or host pointer to input scalar that is overwritten.
[in]y1device pointer or host pointer to input scalar.
[out]paramdevice vector or host vector of 5 elements defining the rotation. param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

◆ hipblasSrotmgBatched()

hipblasStatus_t hipblasSrotmgBatched ( hipblasHandle_t  handle,
float *const  d1[],
float *const  d2[],
float *const  x1[],
const float *const  y1[],
float *const  param[],
int  batchCount 
)

BLAS Level 1 API.

rotmgBatched creates the modified Givens rotation matrix for the batched vectors (d1_i * x1_i, d2_i * y1_i), for i = 1, ..., batchCount. Parameters may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]d1device batched array or host batched array of input scalars that is overwritten.
[in,out]d2device batched array or host batched array of input scalars that is overwritten.
[in,out]x1device batched array or host batched array of input scalars that is overwritten.
[in]y1device batched array or host batched array of input scalars.
[out]paramdevice batched array or host batched array of vectors of 5 elements defining the rotation. param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.
[in]batchCount[int] the number of instances in the batch.

◆ hipblasSrotmgStridedBatched()

hipblasStatus_t hipblasSrotmgStridedBatched ( hipblasHandle_t  handle,
float *  d1,
hipblasStride  strided1,
float *  d2,
hipblasStride  strided2,
float *  x1,
hipblasStride  stridex1,
const float *  y1,
hipblasStride  stridey1,
float *  param,
hipblasStride  strideParam,
int  batchCount 
)

BLAS Level 1 API.

rotmgStridedBatched creates the modified Givens rotation matrix for the strided batched vectors (d1_i * x1_i, d2_i * y1_i), for i = 1, ..., batchCount. Parameters may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]d1device strided_batched array or host strided_batched array of input scalars that is overwritten.
[in]strided1[hipblasStride] specifies the increment between the beginning of d1_i and d1_(i+1)
[in,out]d2device strided_batched array or host strided_batched array of input scalars that is overwritten.
[in]strided2[hipblasStride] specifies the increment between the beginning of d2_i and d2_(i+1)
[in,out]x1device strided_batched array or host strided_batched array of input scalars that is overwritten.
[in]stridex1[hipblasStride] specifies the increment between the beginning of x1_i and x1_(i+1)
[in]y1device strided_batched array or host strided_batched array of input scalars.
[in]stridey1[hipblasStride] specifies the increment between the beginning of y1_i and y1_(i+1)
[out]paramdevice stridedBatched array or host stridedBatched array of vectors of 5 elements defining the rotation. param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.
[in]strideParam[hipblasStride] specifies the increment between the beginning of param_i and param_(i + 1)
[in]batchCount[int] the number of instances in the batch.

◆ hipblasSrotmStridedBatched()

hipblasStatus_t hipblasSrotmStridedBatched ( hipblasHandle_t  handle,
int  n,
float *  x,
int  incx,
hipblasStride  stridex,
float *  y,
int  incy,
hipblasStride  stridey,
const float *  param,
hipblasStride  strideParam,
int  batchCount 
)

BLAS Level 1 API.

rotmStridedBatched applies the modified Givens rotation matrix defined by param_i to strided batched vectors x_i and y_i, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice pointer pointing to first strided batched vector x_1.
[in]incx[int] specifies the increment between elements of each x_i.
[in]stridex[hipblasStride] specifies the increment between the beginning of x_i and x_(i + 1)
[in,out]ydevice pointer pointing to first strided batched vector y_1.
[in]incy[int] specifies the increment between elements of each y_i.
[in]stridey[hipblasStride] specifies the increment between the beginning of y_i and y_(i + 1)
[in]paramdevice pointer pointing to first array of 5 elements defining the rotation (param_1). param[0] = flag param[1] = H11 param[2] = H21 param[3] = H12 param[4] = H22 The flag parameter defines the form of H: flag = -1 => H = ( H11 H12 H21 H22 ) flag = 0 => H = ( 1.0 H12 H21 1.0 ) flag = 1 => H = ( H11 1.0 -1.0 H22 ) flag = -2 => H = ( 1.0 0.0 0.0 1.0 ) param may ONLY be stored on the device for the strided_batched version of this function.
[in]strideParam[hipblasStride] specifies the increment between the beginning of param_i and param_(i + 1)
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasSrotStridedBatched()

hipblasStatus_t hipblasSrotStridedBatched ( hipblasHandle_t  handle,
int  n,
float *  x,
int  incx,
hipblasStride  stridex,
float *  y,
int  incy,
hipblasStride  stridey,
const float *  c,
const float *  s,
int  batchCount 
)

BLAS Level 1 API.

rotStridedBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to strided batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment between elements of each x_i.
[in]stridex[hipblasStride] specifies the increment from the beginning of x_i to the beginning of x_(i+1)
[in,out]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment between elements of each y_i.
[in]stridey[hipblasStride] specifies the increment from the beginning of y_i to the beginning of y_(i+1)
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasSsbmv()

hipblasStatus_t hipblasSsbmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const float *  alpha,
const float *  AP,
int  lda,
const float *  x,
int  incx,
const float *  beta,
float *  y,
int  incy 
)

BLAS Level 2 API.

sbmv performs the matrix-vector operation:

y := alpha*A*x + beta*y,

where alpha and beta are scalars, x and y are n element vectors and A should contain an upper or lower triangular n by n symmetric banded matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : s,d
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int]
[in]k[int] specifies the number of sub- and super-diagonals
[in]alphaspecifies the scalar alpha
[in]APpointer storing matrix A on the GPU
[in]lda[int] specifies the leading dimension of matrix A
[in]xpointer storing vector x on the GPU
[in]incx[int] specifies the increment for the elements of x
[in]betaspecifies the scalar beta
[out]ypointer storing vector y on the GPU
[in]incy[int] specifies the increment for the elements of y

◆ hipblasSsbmvBatched()

hipblasStatus_t hipblasSsbmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const float *  alpha,
const float *const  AP[],
int  lda,
const float *const  x[],
int  incx,
const float *  beta,
float *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

sbmvBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric banded matrix, for i = 1, ..., batchCount. A should contain an upper or lower triangular n by n symmetric banded matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]k[int] specifies the number of sub- and super-diagonals
[in]alphadevice pointer or host pointer to scalar alpha
[in]APdevice array of device pointers storing each matrix A_i
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]xdevice array of device pointers storing each vector x_i
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]betadevice pointer or host pointer to scalar beta
[out]ydevice array of device pointers storing each vector y_i
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]batchCount[int] number of instances in the batch

◆ hipblasSsbmvStridedBatched()

hipblasStatus_t hipblasSsbmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
const float *  x,
int  incx,
hipblasStride  stridex,
const float *  beta,
float *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

sbmvStridedBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric banded matrix, for i = 1, ..., batchCount. A should contain an upper or lower triangular n by n symmetric banded matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]k[int] specifies the number of sub- and super-diagonals
[in]alphadevice pointer or host pointer to scalar alpha
[in]APDevice pointer to the first matrix A_1 on the GPU
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xDevice pointer to the first vector x_1 on the GPU
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. This typically means stridex >= n * incx. stridex should be non zero.
[in]betadevice pointer or host pointer to scalar beta
[out]yDevice pointer to the first vector y_1 on the GPU
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. This typically means stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasSscal()

hipblasStatus_t hipblasSscal ( hipblasHandle_t  handle,
int  n,
const float *  alpha,
float *  x,
int  incx 
)

BLAS Level 1 API.

scal scales each element of vector x with scalar alpha.

x := alpha * x
  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]alphadevice pointer or host pointer for the scalar alpha.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasSscalBatched()

hipblasStatus_t hipblasSscalBatched ( hipblasHandle_t  handle,
int  n,
const float *  alpha,
float *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 1 API.

scalBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasSscalStridedBatched()

hipblasStatus_t hipblasSscalStridedBatched ( hipblasHandle_t  handle,
int  n,
const float *  alpha,
float *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 1 API.

scalStridedBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i ,

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasSspmv()

hipblasStatus_t hipblasSspmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  AP,
const float *  x,
int  incx,
const float *  beta,
float *  y,
int  incy 
)

BLAS Level 2 API.

spmv performs the matrix-vector operation:

y := alpha*A*x + beta*y,

where alpha and beta are scalars, x and y are n element vectors and A should contain an upper or lower triangular n by n packed symmetric matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : s,d
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int]
[in]alphaspecifies the scalar alpha
[in]APpointer storing matrix A on the GPU
[in]xpointer storing vector x on the GPU
[in]incx[int] specifies the increment for the elements of x
[in]betaspecifies the scalar beta
[out]ypointer storing vector y on the GPU
[in]incy[int] specifies the increment for the elements of y

◆ hipblasSspmvBatched()

hipblasStatus_t hipblasSspmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *const  AP[],
const float *const  x[],
int  incx,
const float *  beta,
float *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

spmvBatched performs the matrix-vector operation:

y_i := alpha*AP_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A should contain an upper or lower triangular n by n packed symmetric matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APdevice array of device pointers storing each matrix A_i
[in]xdevice array of device pointers storing each vector x_i
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]betadevice pointer or host pointer to scalar beta
[out]ydevice array of device pointers storing each vector y_i
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]batchCount[int] number of instances in the batch

◆ hipblasSspmvStridedBatched()

hipblasStatus_t hipblasSspmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  AP,
hipblasStride  strideA,
const float *  x,
int  incx,
hipblasStride  stridex,
const float *  beta,
float *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

spmvStridedBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A should contain an upper or lower triangular n by n packed symmetric matrix.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APDevice pointer to the first matrix A_1 on the GPU
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xDevice pointer to the first vector x_1 on the GPU
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. This typically means stridex >= n * incx. stridex should be non zero.
[in]betadevice pointer or host pointer to scalar beta
[out]yDevice pointer to the first vector y_1 on the GPU
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. This typically means stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasSspr()

hipblasStatus_t hipblasSspr ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  x,
int  incx,
float *  AP 
)

BLAS Level 2 API.

spr performs the matrix-vector operations

A := A + alpha*x*x**T

where alpha is a scalar, x is a vector, and A is an n by n symmetric matrix, supplied in packed form.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in AP.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing the packed version of the specified triangular portion of the symmetric matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0

◆ hipblasSspr2()

hipblasStatus_t hipblasSspr2 ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  x,
int  incx,
const float *  y,
int  incy,
float *  AP 
)

BLAS Level 2 API.

spr2 performs the matrix-vector operation

A := A + alpha*x*y**T + alpha*y*x**T

where alpha is a scalar, x and y are vectors, and A is an n by n symmetric matrix, supplied in packed form.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : s,d
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in AP.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing the packed version of the specified triangular portion of the symmetric matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(n) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0

◆ hipblasSspr2Batched()

hipblasStatus_t hipblasSspr2Batched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *const  x[],
int  incx,
const float *const  y[],
int  incy,
float *const  AP[],
int  batchCount 
)

BLAS Level 2 API.

spr2Batched performs the matrix-vector operation

A_i := A_i + alpha*x_i*y_i**T + alpha*y_i*x_i**T

where alpha is a scalar, x_i and y_i are vectors, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each symmetric matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(n) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]batchCount[int] number of instances in the batch.

◆ hipblasSspr2StridedBatched()

hipblasStatus_t hipblasSspr2StridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  x,
int  incx,
hipblasStride  stridex,
const float *  y,
int  incy,
hipblasStride  stridey,
float *  AP,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

spr2StridedBatched performs the matrix-vector operation

A_i := A_i + alpha*x_i*y_i**T + alpha*y_i*x_i**T

where alpha is a scalar, x_i amd y_i are vectors, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in]ydevice pointer pointing to the first vector (y_1).
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1).
[in,out]APdevice pointer storing the packed version of the specified triangular portion of each symmetric matrix A_i. Points to the first A_1. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(n) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasSsprBatched()

hipblasStatus_t hipblasSsprBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *const  x[],
int  incx,
float *const  AP[],
int  batchCount 
)

BLAS Level 2 API.

sprBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**T

where alpha is a scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each symmetric matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]batchCount[int] number of instances in the batch.

◆ hipblasSsprStridedBatched()

hipblasStatus_t hipblasSsprStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  x,
int  incx,
hipblasStride  stridex,
float *  AP,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

sprStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**T

where alpha is a scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in,out]APdevice pointer storing the packed version of the specified triangular portion of each symmetric matrix A_i. Points to the first A_1. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasSswap()

hipblasStatus_t hipblasSswap ( hipblasHandle_t  handle,
int  n,
float *  x,
int  incx,
float *  y,
int  incy 
)

BLAS Level 1 API.

swap interchanges vectors x and y.

y := x; x := y
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasSswapBatched()

hipblasStatus_t hipblasSswapBatched ( hipblasHandle_t  handle,
int  n,
float *const  x[],
int  incx,
float *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

swapBatched interchanges vectors x_i and y_i, for i = 1 , ... , batchCount

y_i := x_i; x_i := y_i
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasSswapStridedBatched()

hipblasStatus_t hipblasSswapStridedBatched ( hipblasHandle_t  handle,
int  n,
float *  x,
int  incx,
hipblasStride  stridex,
float *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

swapStridedBatched interchanges vectors x_i and y_i, for i = 1 , ... , batchCount

y_i := x_i; x_i := y_i
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in,out]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in,out]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_y is of appropriate size, for a typical case this means stride_y >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch.

◆ hipblasSsymm()

hipblasStatus_t hipblasSsymm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
const float *  BP,
int  ldb,
const float *  beta,
float *  CP,
int  ldc 
)

BLAS Level 3 API.

symm performs one of the matrix-matrix operations:

C := alpha*A*B + beta*C if side == HIPBLAS_SIDE_LEFT, C := alpha*B*A + beta*C if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B and C are m by n matrices, and A is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C := alpha*A*B + beta*C HIPBLAS_SIDE_RIGHT: C := alpha*B*A + beta*C
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix
[in]m[int] m specifies the number of rows of B and C. m >= 0.
[in]n[int] n specifies the number of columns of B and C. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A and B are not referenced.
[in]APpointer storing matrix A on the GPU. A is m by m if side == HIPBLAS_SIDE_LEFT A is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPpointer storing matrix B on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m )

◆ hipblasSsymmBatched()

hipblasStatus_t hipblasSsymmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const float *  alpha,
const float *const  AP[],
int  lda,
const float *const  BP[],
int  ldb,
const float *  beta,
float *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

symmBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]m[int] m specifies the number of rows of B_i and C_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i and C_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPdevice array of device pointers storing each matrix B_i on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C_i need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C_i. ldc >= max( 1, m )
[in]batchCount[int] number of instances in the batch.

◆ hipblasSsymmStridedBatched()

hipblasStatus_t hipblasSsymmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
const float *  BP,
int  ldb,
hipblasStride  strideB,
const float *  beta,
float *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

symmStridedBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]m[int] m specifies the number of rows of B_i and C_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i and C_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice pointer to first matrix A_1 A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPdevice pointer to first matrix B_1 of dimension (ldb, n) on the GPU.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m )
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice pointer to first matrix C_1 of dimension (ldc, n) on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasSsymv()

hipblasStatus_t hipblasSsymv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
const float *  x,
int  incx,
const float *  beta,
float *  y,
int  incy 
)

BLAS Level 2 API.

symv performs the matrix-vector operation:

y := alpha*A*x + beta*y,

where alpha and beta are scalars, x and y are n element vectors and A should contain an upper or lower triangular n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int]
[in]alphaspecifies the scalar alpha
[in]APpointer storing matrix A on the GPU
[in]lda[int] specifies the leading dimension of A
[in]xpointer storing vector x on the GPU
[in]incx[int] specifies the increment for the elements of x
[in]betaspecifies the scalar beta
[out]ypointer storing vector y on the GPU
[in]incy[int] specifies the increment for the elements of y

◆ hipblasSsymvBatched()

hipblasStatus_t hipblasSsymvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *const  AP[],
int  lda,
const float *const  x[],
int  incx,
const float *  beta,
float *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

symvBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A a should contain an upper or lower triangular symmetric matrix and the opposing triangular part of A is not referenced

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APdevice array of device pointers storing each matrix A_i
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]xdevice array of device pointers storing each vector x_i
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]betadevice pointer or host pointer to scalar beta
[out]ydevice array of device pointers storing each vector y_i
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]batchCount[int] number of instances in the batch

◆ hipblasSsymvStridedBatched()

hipblasStatus_t hipblasSsymvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
const float *  x,
int  incx,
hipblasStride  stridex,
const float *  beta,
float *  y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

symvStridedBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A a should contain an upper or lower triangular symmetric matrix and the opposing triangular part of A is not referenced

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APDevice pointer to the first matrix A_1 on the GPU
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xDevice pointer to the first vector x_1 on the GPU
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. This typically means stridex >= n * incx. stridex should be non zero.
[in]betadevice pointer or host pointer to scalar beta
[out]yDevice pointer to the first vector y_1 on the GPU
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. This typically means stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasSsyr()

hipblasStatus_t hipblasSsyr ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  x,
int  incx,
float *  AP,
int  lda 
)

BLAS Level 2 API.

syr performs the matrix-vector operations

A := A + alpha*x*x**T

where alpha is a scalar, x is a vector, and A is an n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasSsyr2()

hipblasStatus_t hipblasSsyr2 ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  x,
int  incx,
const float *  y,
int  incy,
float *  AP,
int  lda 
)

BLAS Level 2 API.

syr2 performs the matrix-vector operations

A := A + alpha*x*y**T + alpha*y*x**T

where alpha is a scalar, x and y are vectors, and A is an n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasSsyr2Batched()

hipblasStatus_t hipblasSsyr2Batched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *const  x[],
int  incx,
const float *const  y[],
int  incy,
float *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

syr2Batched performs a batch of matrix-vector operations

A[i] := A[i] + alpha*x[i]*y[i]**T + alpha*y[i]*x[i]**T

where alpha is a scalar, x[i] and y[i] are vectors, and A[i] is a n by n symmetric matrix, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasSsyr2k()

hipblasStatus_t hipblasSsyr2k ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const float *  AP,
int  lda,
const float *  BP,
int  ldb,
const float *  beta,
float *  CP,
int  ldc 
)

BLAS Level 3 API.

syr2k performs one of the matrix-matrix operations for a symmetric rank-2k update

C := alpha*(op( A )*op( B )^T + op( B )*op( A )^T) + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^T, op( B ) = B^T,  and A and B are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A ) = A^T, op( B ) = B^T HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A) and op(B). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasSsyr2kBatched()

hipblasStatus_t hipblasSsyr2kBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const float *const  AP[],
int  lda,
const float *const  BP[],
int  ldb,
const float *  beta,
float *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syr2kBatched performs a batch of the matrix-matrix operations for a symmetric rank-2k update

C_i := alpha*(op( A_i )*op( B_i )^T + op( B_i )*op( A_i )^T) + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasSsyr2kStridedBatched()

hipblasStatus_t hipblasSsyr2kStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
const float *  BP,
int  ldb,
hipblasStride  strideB,
const float *  beta,
float *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syr2kStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-2k update

C_i := alpha*(op( A_i )*op( B_i )^T + op( B_i )*op( A_i )^T) + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasSsyr2StridedBatched()

hipblasStatus_t hipblasSsyr2StridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  x,
int  incx,
hipblasStride  stridex,
const float *  y,
int  incy,
hipblasStride  stridey,
float *  AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

syr2StridedBatched the matrix-vector operations

A[i] := A[i] + alpha*x[i]*y[i]**T + alpha*y[i]*x[i]**T

where alpha is a scalar, x[i] and y[i] are vectors, and A[i] is a n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of each matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the pointer increment between vectors (x_i) and (x_i+1).
[in]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] specifies the pointer increment between vectors (y_i) and (y_i+1).
[in,out]APdevice pointer to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasSsyrBatched()

hipblasStatus_t hipblasSsyrBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *const  x[],
int  incx,
float *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

syrBatched performs a batch of matrix-vector operations

A[i] := A[i] + alpha*x[i]*x[i]**T

where alpha is a scalar, x is an array of vectors, and A is an array of n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasSsyrk()

hipblasStatus_t hipblasSsyrk ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const float *  AP,
int  lda,
const float *  beta,
float *  CP,
int  ldc 
)

BLAS Level 3 API.

syrk performs one of the matrix-matrix operations for a symmetric rank-k update

C := alpha*op( A )*op( A )^T + beta*C

where alpha and beta are scalars, op(A) is an n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower.

op( A ) = A, and A is n by k if transA == HIPBLAS_OP_N
op( A ) = A^T and A is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if transA = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasSsyrkBatched()

hipblasStatus_t hipblasSsyrkBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const float *const  AP[],
int  lda,
const float *  beta,
float *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syrkBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( A_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) is an n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^T and A_i is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasSsyrkStridedBatched()

hipblasStatus_t hipblasSsyrkStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
const float *  beta,
float *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syrkStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( A_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) is an n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^T and A_i is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU. on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasSsyrkx()

hipblasStatus_t hipblasSsyrkx ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const float *  AP,
int  lda,
const float *  BP,
int  ldb,
const float *  beta,
float *  CP,
int  ldc 
)

BLAS Level 3 API.

syrkx performs one of the matrix-matrix operations for a symmetric rank-k update

C := alpha*op( A )*op( B )^T + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A )*op( B )^T will be symmetric.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^T, op( B ) = B^T,  and A and B are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A ) = A^T, op( B ) = B^T HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A) and op(B). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasSsyrkxBatched()

hipblasStatus_t hipblasSsyrkxBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const float *const  AP[],
int  lda,
const float *const  BP[],
int  ldb,
const float *  beta,
float *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syrkxBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( B_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A_i )*op( B_i )^T will be symmetric.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasSsyrkxStridedBatched()

hipblasStatus_t hipblasSsyrkxStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
const float *  BP,
int  ldb,
hipblasStride  strideB,
const float *  beta,
float *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syrkxStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( B_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A_i )*op( B_i )^T will be symmetric.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasSsyrStridedBatched()

hipblasStatus_t hipblasSsyrStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const float *  alpha,
const float *  x,
int  incx,
hipblasStride  stridex,
float *  AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

syrStridedBatched performs the matrix-vector operations

A[i] := A[i] + alpha*x[i]*x[i]**T

where alpha is a scalar, vectors, and A is an array of n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of each matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the pointer increment between vectors (x_i) and (x_i+1).
[in,out]APdevice pointer to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasStatusToString()

const char* hipblasStatusToString ( hipblasStatus_t  status)

HIPBLAS Auxiliary API

hipblasStatusToString

Returns string representing hipblasStatus_t value

Parameters
[in]status[hipblasStatus_t] hipBLAS status to convert to string

◆ hipblasStbmv()

hipblasStatus_t hipblasStbmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const float *  AP,
int  lda,
float *  x,
int  incx 
)

BLAS Level 2 API.

tbmv performs one of the matrix-vector operations

x := A*x      or
x := A**T*x   or
x := A**H*x,

x is a vectors and A is a banded m by m matrix (see description below).

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of A is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of A's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by A.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of the matrix A. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of the matrix A. k must satisfy k > 0 && k < lda.
[in]APdevice pointer storing banded triangular matrix A. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of A. lda must satisfy lda > k.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasStbmvBatched()

hipblasStatus_t hipblasStbmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const float *const  AP[],
int  lda,
float *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tbmvBatched performs one of the matrix-vector operations

x_i := A_i*x_i      or
x_i := A_i**T*x_i   or
x_i := A_i**H*x_i,

where (A_i, x_i) is the i-th instance of the batch. x_i is a vector and A_i is an m by m matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether each matrix A_i is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of each A_i is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of each A_i's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by each A_i.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of each matrix A_i. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of each matrix A_i. k must satisfy k > 0 && k < lda.
[in]APdevice array of device pointers storing each banded triangular matrix A_i. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of each A_i. lda must satisfy lda > k.
[in,out]xdevice array of device pointer storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasStbmvStridedBatched()

hipblasStatus_t hipblasStbmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const float *  AP,
int  lda,
hipblasStride  strideA,
float *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tbmvStridedBatched performs one of the matrix-vector operations

x_i := A_i*x_i      or
x_i := A_i**T*x_i   or
x_i := A_i**H*x_i,

where (A_i, x_i) is the i-th instance of the batch. x_i is a vector and A_i is an m by m matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether each matrix A_i is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of each A_i is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of each A_i's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by each A_i.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of each matrix A_i. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of each matrix A_i. k must satisfy k > 0 && k < lda.
[in]APdevice array to the first matrix A_i of the batch. Stores each banded triangular matrix A_i. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of each A_i. lda must satisfy lda > k.
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in,out]xdevice array to the first vector x_i of the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one x_i matrix to the next x_(i + 1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasStbsv()

hipblasStatus_t hipblasStbsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const float *  AP,
int  lda,
float *  x,
int  incx 
)

BLAS Level 2 API.

tbsv solves

 A*x = b or A**T*x = b or A**H*x = b,

where x and b are vectors and A is a banded triangular matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular (i.e. the diagonal elements of A are not used in computations). HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of b. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of A. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of A. k >= 0.
[in]APdevice pointer storing the matrix A in banded format.
[in]lda[int] specifies the leading dimension of A. lda >= (k + 1).
[in,out]xdevice pointer storing input vector b. Overwritten by the output vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasStbsvBatched()

hipblasStatus_t hipblasStbsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const float *const  AP[],
int  lda,
float *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tbsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a banded triangular matrix, for i = [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A_i*x_i = b_i HIPBLAS_OP_T: Solves A_i**T*x_i = b_i HIPBLAS_OP_C: Solves A_i**H*x_i = b_i
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of each b_i. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of each A_i. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of each A_i. k >= 0.
[in]APdevice vector of device pointers storing each matrix A_i in banded format.
[in]lda[int] specifies the leading dimension of each A_i. lda >= (k + 1).
[in,out]xdevice vector of device pointers storing each input vector b_i. Overwritten by each output vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasStbsvStridedBatched()

hipblasStatus_t hipblasStbsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const float *  AP,
int  lda,
hipblasStride  strideA,
float *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tbsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a banded triangular matrix, for i = [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A_i*x_i = b_i HIPBLAS_OP_T: Solves A_i**T*x_i = b_i HIPBLAS_OP_C: Solves A_i**H*x_i = b_i
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of each b_i. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of each A_i. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of each A_i. k >= 0.
[in]APdevice pointer pointing to the first banded matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i. lda >= (k + 1).
[in]strideA[hipblasStride] specifies the distance between the start of one matrix (A_i) and the next (A_i+1).
[in,out]xdevice pointer pointing to the first input vector b_1. Overwritten by output vectors x.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the distance between the start of one vector (x_i) and the next (x_i+1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasStpmv()

hipblasStatus_t hipblasStpmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *  AP,
float *  x,
int  incx 
)

BLAS Level 2 API.

tpmv performs one of the matrix-vector operations

 x = A*x or x = A**T*x,

where x is an n element vector and A is an n by n unit, or non-unit, upper or lower triangular matrix, supplied in the pack form.

The vector x is overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of A. m >= 0.
[in]APdevice pointer storing matrix A, of dimension at leat ( m * ( m + 1 ) / 2 ). Before entry with uplo = HIPBLAS_FILL_MODE_UPPER, the array A must contain the upper triangular matrix packed sequentially, column by column, so that A[0] contains a_{0,0}, A[1] and A[2] contain a_{0,1} and a_{1, 1} respectively, and so on. Before entry with uplo = HIPBLAS_FILL_MODE_LOWER, the array A must contain the lower triangular matrix packed sequentially, column by column, so that A[0] contains a_{0,0}, A[1] and A[2] contain a_{1,0} and a_{2,0} respectively, and so on. Note that when DIAG = HIPBLAS_DIAG_UNIT, the diagonal elements of A are not referenced, but are assumed to be unity.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x. incx must not be zero.

◆ hipblasStpmvBatched()

hipblasStatus_t hipblasStpmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *const  AP[],
float *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tpmvBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix)

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer storing pointer of matrices A_i, of dimension ( lda, m )
[in]xdevice pointer storing vectors x_i.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasStpmvStridedBatched()

hipblasStatus_t hipblasStpmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *  AP,
hipblasStride  strideA,
float *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tpmvStridedBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix) with strides specifying how to retrieve $x_i$ (resp. $A_i$) from $x_{i-1}$ (resp. $A_i$).

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer of the matrix A_0, of dimension ( lda, m )
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_{i + 1}
[in]xdevice pointer storing the vector x_0.
[in]incx[int] specifies the increment for the elements of one vector x.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_{i + 1}
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasStpsv()

hipblasStatus_t hipblasStpsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *  AP,
float *  x,
int  incx 
)

BLAS Level 2 API.

tpsv solves

 A*x = b or A**T*x = b, or A**H*x = b,

where x and b are vectors and A is a triangular matrix stored in the packed format.

The input vector b is overwritten by the output vector x.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular (i.e. the diagonal elements of A are not used in computations). HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice pointer storing the packed version of matrix A, of dimension >= (n * (n + 1) / 2)
[in,out]xdevice pointer storing vector b on input, overwritten by x on output.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasStpsvBatched()

hipblasStatus_t hipblasStpsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *const  AP[],
float *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tpsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i, or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a triangular matrix stored in the packed format, for i in [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice array of device pointers storing the packed versions of each matrix A_i, of dimension >= (n * (n + 1) / 2)
[in,out]xdevice array of device pointers storing each input vector b_i, overwritten by x_i on output.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasStpsvStridedBatched()

hipblasStatus_t hipblasStpsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *  AP,
hipblasStride  strideA,
float *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tpsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i, or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a triangular matrix stored in the packed format, for i in [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice pointer pointing to the first packed matrix A_1, of dimension >= (n * (n + 1) / 2)
[in]strideA[hipblasStride] stride from the beginning of one packed matrix (AP_i) and the next (AP_i+1).
[in,out]xdevice pointer pointing to the first input vector b_1. Overwritten by each x_i on output.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the beginning of one vector (x_i) and the next (x_i+1).
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasStrmm()

hipblasStatus_t hipblasStrmm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
float *  BP,
int  ldb 
)

BLAS Level 3 API.

#ifndef HIPBLAS_V1 // deprecated

    hipblasStatus_t hipblasStrmm(hipblasHandle_t    handle,
                                 hipblasSideMode_t  side,
                                 hipblasFillMode_t  uplo,
                                 hipblasOperation_t transA,
                                 hipblasDiagType_t  diag,
                                 int                m,
                                 int                n,
                                 const float*       alpha,
                                 const float*       AP,
                                 int                lda,
                                 float*             BP,
                                 int                ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasStrmmOutofplace(hipblasHandle_t    handle,
                                          hipblasSideMode_t  side,
                                          hipblasFillMode_t  uplo,
                                          hipblasOperation_t transA,
                                          hipblasDiagType_t  diag,
                                          int                m,
                                          int                n,
                                          const float*       alpha,
                                          const float*       AP,
                                          int                lda,
                                          const float*       BP,
                                          int                ldb,
                                          float*             CP,
                                          int                ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasDtrmm(hipblasHandle_t    handle,
                                hipblasSideMode_t  side,
                                hipblasFillMode_t  uplo,
                                hipblasOperation_t transA,
                                hipblasDiagType_t  diag,
                                int                m,
                                int                n,
                                const double*      alpha,
                                const double*      AP,
                                int                lda,
                                double*            BP,
                                int                ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasDtrmmOutofplace(hipblasHandle_t    handle,
                                          hipblasSideMode_t  side,
                                          hipblasFillMode_t  uplo,
                                          hipblasOperation_t transA,
                                          hipblasDiagType_t  diag,
                                          int                m,
                                          int                n,
                                          const double*      alpha,
                                          const double*      AP,
                                          int                lda,
                                          const double*      BP,
                                          int                ldb,
                                          double*            CP,
                                          int                ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasCtrmm(hipblasHandle_t       handle,
                                hipblasSideMode_t     side,
                                hipblasFillMode_t     uplo,
                                hipblasOperation_t    transA,
                                hipblasDiagType_t     diag,
                                int                   m,
                                int                   n,
                                const hipblasComplex* alpha,
                                const hipblasComplex* AP,
                                int                   lda,
                                hipblasComplex*       BP,
                                int                   ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasCtrmmOutofplace(hipblasHandle_t       handle,
                                          hipblasSideMode_t     side,
                                          hipblasFillMode_t     uplo,
                                          hipblasOperation_t    transA,
                                          hipblasDiagType_t     diag,
                                          int                   m,
                                          int                   n,
                                          const hipblasComplex* alpha,
                                          const hipblasComplex* AP,
                                          int                   lda,
                                          const hipblasComplex* BP,
                                          int                   ldb,
                                          hipblasComplex*       CP,
                                          int                   ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasZtrmm(hipblasHandle_t             handle,
                                hipblasSideMode_t           side,
                                hipblasFillMode_t           uplo,
                                hipblasOperation_t          transA,
                                hipblasDiagType_t           diag,
                                int                         m,
                                int                         n,
                                const hipblasDoubleComplex* alpha,
                                const hipblasDoubleComplex* AP,
                                int                         lda,
                                hipblasDoubleComplex*       BP,
                                int                         ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasZtrmmOutofplace(hipblasHandle_t             handle,
                                          hipblasSideMode_t           side,
                                          hipblasFillMode_t           uplo,
                                          hipblasOperation_t          transA,
                                          hipblasDiagType_t           diag,
                                          int                         m,
                                          int                         n,
                                          const hipblasDoubleComplex* alpha,
                                          const hipblasDoubleComplex* AP,
                                          int                         lda,
                                          const hipblasDoubleComplex* BP,
                                          int                         ldb,
                                          hipblasDoubleComplex*       CP,
                                          int                         ldc);

#endif

The deprecated Legacy BLAS in-place trmm performs one of the matrix-matrix operations:

B := alpha*op( A )*B,   or
B := alpha*B*op( A ),

The new trmm performs one of the matrix-matrix operations:

C := alpha*op( A )*B,   or
C := alpha*B*op( A ),

The in-place functionality is still available in the new trmmm by setting pointer C equal to pointer B, and ldc equal to ldb.

alpha  is a scalar,  B  is an m by n matrix, C  is an m by n matrix,  A  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A )  is one  of

op( A ) = A     or
op( A ) = A^T   or
op( A ) = A^H.

When uplo == rocblas_fill_upper the  leading  k by k
upper triangular part of the array  A must contain the upper
triangular matrix and the strictly lower triangular part of
A is not referenced. Here k is m when side == rocblas_side_left
and is n when side == rocblas_side_right.

When uplo == rocblas_fill_lower the  leading  k by k
lower triangular part of the array  A must contain the lower
triangular matrix  and the strictly upper triangular part of
A is not referenced. Here k is m when  side == rocblas_side_left
and is n when side == rocblas_side_right.

Note that when  diag == rocblas_diagonal_unit  the diagonal elements of
A  are not referenced either,  but are assumed to be  unity.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A) multiplies B from the left or right as follows: HIPBLAS_SIDE_LEFT: B := alpha*op( A )*B. HIPBLAS_SIDE_RIGHT: B := alpha*B*op( A ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A is unit triangular as follows: HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B. m >= 0.
[in]n[int] n specifies the number of columns of B. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and B need not be set before entry.
[in]APDevice pointer to matrix A on the GPU. A has dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPDevice pointer to the first matrix B_0 on the GPU. On entry, the leading m by n part of the array B must contain the matrix B, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m ).

◆ hipblasStrmmBatched()

hipblasStatus_t hipblasStrmmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const float *  alpha,
const float *const  AP[],
int  lda,
float *const  BP[],
int  ldb,
int  batchCount 
)

BLAS Level 3 API.

The hipBLAS trmm_batched API is from Legacy BLAS and it supports only in-place functionality. It is deprecated and it will be replaced with an API that supports both in-place and out-of-place functionality. The new API is available in hipBLAS versions 1.x.x and later. To get the new API compile with the directive -DHIPBLAS_V1.

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasStrmmBatched(hipblasHandle_t    handle,
                                        hipblasSideMode_t  side,
                                        hipblasFillMode_t  uplo,
                                        hipblasOperation_t transA,
                                        hipblasDiagType_t  diag,
                                        int                m,
                                        int                n,
                                        const float*       alpha,
                                        const float* const AP[],
                                        int                lda,
                                        float* const       BP[],
                                        int                ldb,
                                        int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasStrmmBatchedOutofplace(
                                        hipblasHandle_t    handle,
                                        hipblasSideMode_t  side,
                                        hipblasFillMode_t  uplo,
                                        hipblasOperation_t transA,
                                        hipblasDiagType_t  diag,
                                        int                m,
                                        int                n,
                                        const float*       alpha,
                                        const float* const AP[],
                                        int                lda,
                                        const float* const BP[],
                                        int                ldb,
                                        float* const       CP[],
                                        int                ldc,
                                        int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasDtrmmBatched(hipblasHandle_t     handle,
                                        hipblasSideMode_t   side,
                                        hipblasFillMode_t   uplo,
                                        hipblasOperation_t  transA,
                                        hipblasDiagType_t   diag,
                                        int                 m,
                                        int                 n,
                                        const double*       alpha,
                                        const double* const AP[],
                                        int                 lda,
                                        double* const       BP[],
                                        int                 ldb,
                                        int                 batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasDtrmmBatchedOutofplace(
                                        hipblasHandle_t     handle,
                                        hipblasSideMode_t   side,
                                        hipblasFillMode_t   uplo,
                                        hipblasOperation_t  transA,
                                        hipblasDiagType_t   diag,
                                        int                 m,
                                        int                 n,
                                        const double*       alpha,
                                        const double* const AP[],
                                        int                 lda,
                                        const double* const BP[],
                                        int                 ldb,
                                        double* const       CP[],
                                        int                 ldc,
                                        int                 batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasCtrmmBatched(hipblasHandle_t             handle,
                                        hipblasSideMode_t           side,
                                        hipblasFillMode_t           uplo,
                                        hipblasOperation_t          transA,
                                        hipblasDiagType_t           diag,
                                        int                         m,
                                        int                         n,
                                        const hipblasComplex*       alpha,
                                        const hipblasComplex* const AP[],
                                        int                         lda,
                                        hipblasComplex* const       BP[],
                                        int                         ldb,
                                        int                         batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasCtrmmBatchedOutofplace(
                                        hipblasHandle_t             handle,
                                        hipblasSideMode_t           side,
                                        hipblasFillMode_t           uplo,
                                        hipblasOperation_t          transA,
                                        hipblasDiagType_t           diag,
                                        int                         m,
                                        int                         n,
                                        const hipblasComplex*       alpha,
                                        const hipblasComplex* const AP[],
                                        int                         lda,
                                        const hipblasComplex* const BP[],
                                        int                         ldb,
                                        hipblasComplex* const       CP[],
                                        int                         ldc,
                                        int batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasZtrmmBatched(hipblasHandle_t                   handle,
                                        hipblasSideMode_t                 side,
                                        hipblasFillMode_t                 uplo,
                                        hipblasOperation_t                transA,
                                        hipblasDiagType_t                 diag,
                                        int                               m,
                                        int                               n,
                                        const hipblasDoubleComplex*       alpha,
                                        const hipblasDoubleComplex* const AP[],
                                        int                               lda,
                                        hipblasDoubleComplex* const       BP[],
                                        int                               ldb,
                                        int                               batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasZtrmmBatchedOutofplace(
                                        hipblasHandle_t                   handle,
                                        hipblasSideMode_t                 side,
                                        hipblasFillMode_t                 uplo,
                                        hipblasOperation_t                transA,
                                        hipblasDiagType_t                 diag,
                                        int                               m,
                                        int                               n,
                                        const hipblasDoubleComplex*       alpha,
                                        const hipblasDoubleComplex* const AP[],
                                        int                               lda,
                                        const hipblasDoubleComplex* const BP[],
                                        int                               ldb,
                                        hipblasDoubleComplex* const       CP[],
                                        int                               ldc,
                                        int                               batchCount);

#endif

The deprecated Legacy BLAS in-place trmm_batched performs one of the batched matrix-matrix operations:

B_i := alpha*op( A_i )*B_i,   or
B_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The new trmm_batched performs one of the matrix-matrix operations:

C_i := alpha*op( A_i )*B_i,   or
C_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The in-place functionality is still available in the new trmmm_batched by setting pointer C equal to pointer B and ldc equal to ldb.

alpha  is a scalar,  B_i  is an m by n matrix, C_i  is an m by n matrix,  A_i  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A_i )  is one  of

op( A_i ) = A_i   or   op( A_i ) = A_i^T   or   op( A_i ) = A_i^H.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A_i) multiplies B_i from the left or right as follows: HIPBLAS_SIDE_LEFT: B_i := alpha*op( A_i )*B_i. HIPBLAS_SIDE_RIGHT: B_i := alpha*B_i*op( A_i ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A_i) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A_i) = A_i. HIPBLAS_OP_T: op(A_i) = A_i^T. HIPBLAS_OP_C: op(A_i) = A_i^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A_i is unit triangular as follows: HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i is not referenced and B_i need not be set before entry.
[in]APDevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A_i are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice array of device pointers storing each matrix B_i on the GPU. On entry, the leading m by n part of the array B_i must contain the matrix B_i, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m ).
[in]batchCount[int] number of instances i in the batch.

◆ hipblasStrmmBatchedOutofplace()

hipblasStatus_t hipblasStrmmBatchedOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const float *  alpha,
const float *const  AP[],
int  lda,
const float *const  BP[],
int  ldb,
float *const  CP[],
int  ldc,
int  batchCount 
)

◆ hipblasStrmmOutofplace()

hipblasStatus_t hipblasStrmmOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
const float *  BP,
int  ldb,
float *  CP,
int  ldc 
)

◆ hipblasStrmmStridedBatched()

hipblasStatus_t hipblasStrmmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
float *  BP,
int  ldb,
hipblasStride  strideB,
int  batchCount 
)

BLAS Level 3 API.

The hipBLAS trmm_strided_batched API is from Legacy BLAS and it supports only in-place functionality. It is deprecated and it will be replaced with an API that supports both in-place and out-of-place functionality. The new API is available in hipBLAS versions 1.x.x and later. To get the new API compile with the directive -DHIPBLAS_V1.

#ifndef HIPBLAS_V1  //  deprecated
       hipblasStatus_t hipblasStrmmStridedBatched(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const float*       alpha,
                           const float*       AP,
                           int                lda,
                           hipblasStride      strideA,
                           float*             BP,
                           int                ldb,
                           hipblasStride      strideB,
                           int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasStrmmStridedBatchedOutofplace(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const float*       alpha,
                           const float*       AP,
                           int                lda,
                           hipblasStride      strideA,
                           const float*       BP,
                           int                ldb,
                           hipblasStride      strideB,
                           float*             CP,
                           int                ldc,
                           hipblasStride      strideC,
                           int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated
       hipblasStatus_t hipblasDtrmmStridedBatched(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const double*      alpha,
                           const double*      AP,
                           int                lda,
                           hipblasStride      strideA,
                           double*            BP,
                           int                ldb,
                           hipblasStride      strideB,
                           int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasDtrmmStridedBatchedOutofplace(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const double*      alpha,
                           const double*      AP,
                           int                lda,
                           hipblasStride      strideA,
                           const double*      BP,
                           int                ldb,
                           hipblasStride      strideB,
                           double*            CP,
                           int                ldc,
                           hipblasStride      strideC,
                           int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

       hipblasStatus_t hipblasCtrmmStridedBatched(
                           hipblasHandle_t       handle,
                           hipblasSideMode_t     side,
                           hipblasFillMode_t     uplo,
                           hipblasOperation_t    transA,
                           hipblasDiagType_t     diag,
                           int                   m,
                           int                   n,
                           const hipblasComplex* alpha,
                           const hipblasComplex* AP,
                           int                   lda,
                           hipblasStride         strideA,
                           hipblasComplex*       BP,
                           int                   ldb,
                           hipblasStride         strideB,
                           int                   batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasCtrmmStridedBatchedOutofplace(
                           hipblasHandle_t       handle,
                           hipblasSideMode_t     side,
                           hipblasFillMode_t     uplo,
                           hipblasOperation_t    transA,
                           hipblasDiagType_t     diag,
                           int                   m,
                           int                   n,
                           const hipblasComplex* alpha,
                           const hipblasComplex* AP,
                           int                   lda,
                           hipblasStride         strideA,
                           const hipblasComplex* BP,
                           int                   ldb,
                           hipblasStride         strideB,
                           hipblasComplex*       CP,
                           int                   ldc,
                           hipblasStride         strideC,
                           int batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasZtrmmStridedBatched(hipblasHandle_t             handle,
                                               hipblasSideMode_t           side,
                                               hipblasFillMode_t           uplo,
                                               hipblasOperation_t          transA,
                                               hipblasDiagType_t           diag,
                                               int                         m,
                                               int                         n,
                                               const hipblasDoubleComplex* alpha,
                                               const hipblasDoubleComplex* AP,
                                               int                         lda,
                                               hipblasStride               strideA,
                                               hipblasDoubleComplex*       BP,
                                               int                         ldb,
                                               hipblasStride               strideB,
                                               int                         batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasZtrmmStridedBatchedOutofplace(
                                                hipblasHandle_t             handle,
                                                hipblasSideMode_t           side,
                                                hipblasFillMode_t           uplo,
                                                hipblasOperation_t          transA,
                                                hipblasDiagType_t           diag,
                                                int                         m,
                                                int                         n,
                                                const hipblasDoubleComplex* alpha,
                                                const hipblasDoubleComplex* AP,
                                                int                         lda,
                                                hipblasStride               strideA,
                                                const hipblasDoubleComplex* BP,
                                                int                         ldb,
                                                hipblasStride               strideB,
                                                hipblasDoubleComplex*       BC,
                                                int                         ldc,
                                                hipblasStride               strideC,
                                                int                         batchCount);

#endif

The deprecated Legacy BLAS in-place trmm_strided_batched performs one of the strided_batched matrix-matrix operations:

B_i := alpha*op( A_i )*B_i,   or
B_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The new trmm_batched performs one of the matrix-matrix operations:

C_i := alpha*op( A_i )*B_i,   or
C_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The in-place functionality is still available in the new trmmm_batched by setting pointer C equal to pointer B, setting ldc equal to ldb, and setting stride_C equal to stride_B.

alpha  is a scalar,  B_i  is an m by n matrix, C_i  is an m by n matrix,  A_i  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A_i )  is one  of

op( A_i ) = A_i   or
op( A_i ) = A_i^T   or
op( A_i ) = A_i^H.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A_i) multiplies B_i from the left or right as follows: HIPBLAS_SIDE_LEFT: B_i := alpha*op( A_i )*B_i. HIPBLAS_SIDE_RIGHT: B_i := alpha*B_i*op( A_i ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A_i) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A_i) = A_i. HIPBLAS_OP_T: op(A_i) = A_i^T. HIPBLAS_OP_C: op(A_i) = A_i^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A_i is unit triangular as follows: HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i is not referenced and B_i need not be set before entry.
[in]APDevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A_i are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in,out]BPDevice pointer to the first matrix B_0 on the GPU. On entry, the leading m by n part of the array B_i must contain the matrix B_i, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasStrmmStridedBatchedOutofplace()

hipblasStatus_t hipblasStrmmStridedBatchedOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
const float *  BP,
int  ldb,
hipblasStride  strideB,
float *  CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

◆ hipblasStrmv()

hipblasStatus_t hipblasStrmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *  AP,
int  lda,
float *  x,
int  incx 
)

BLAS Level 2 API.

trmv performs one of the matrix-vector operations

 x = A*x or x = A**T*x,

where x is an n element vector and A is an n by n unit, or non-unit, upper or lower triangular matrix.

The vector x is overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of A. m >= 0.
[in]APdevice pointer storing matrix A, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A. lda = max( 1, m ).
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasStrmvBatched()

hipblasStatus_t hipblasStrmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *const  AP[],
int  lda,
float *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

trmvBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix)

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer storing pointer of matrices A_i, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A_i. lda >= max( 1, m ).
[in]xdevice pointer storing vectors x_i.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasStrmvStridedBatched()

hipblasStatus_t hipblasStrmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *  AP,
int  lda,
hipblasStride  strideA,
float *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

trmvStridedBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix) with strides specifying how to retrieve $x_i$ (resp. $A_i$) from $x_{i-1}$ (resp. $A_i$).

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer of the matrix A_0, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A_i. lda >= max( 1, m ).
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_{i + 1}
[in]xdevice pointer storing the vector x_0.
[in]incx[int] specifies the increment for the elements of one vector x.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_{i + 1}
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasStrsm()

hipblasStatus_t hipblasStrsm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
float *  BP,
int  ldb 
)

BLAS Level 3 API.

trsm solves

op(A)*X = alpha*B or  X*op(A) = alpha*B,

where alpha is a scalar, X and B are m by n matrices, A is triangular matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

The matrix X is overwritten on B.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time.

(where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B. m >= 0.
[in]n[int] n specifies the number of columns of B. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice pointer storing matrix A. of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice pointer storing matrix B.
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m ).

◆ hipblasStrsmBatched()

hipblasStatus_t hipblasStrsmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const float *  alpha,
const float *const  AP[],
int  lda,
float *const  BP[],
int  ldb,
int  batchCount 
)

BLAS Level 3 API.

trsmBatched performs the following batched operation:

op(A_i)*X_i = alpha*B_i or  X_i*op(A_i) = alpha*B_i, for i = 1, ..., batchCount.

where alpha is a scalar, X and B are batched m by n matrices, A is triangular batched matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

Each matrix X_i is overwritten on B_i for i = 1, ..., batchCount.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time. (where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each B_i. m >= 0.
[in]n[int] n specifies the number of columns of each B_i. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Matricies are of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of each A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice array of device pointers storing each matrix B_i on the GPU.
[in]ldb[int] ldb specifies the first dimension of each B_i. ldb >= max( 1, m ).
[in]batchCount[int] number of trsm operatons in the batch.

◆ hipblasStrsmStridedBatched()

hipblasStatus_t hipblasStrsmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const float *  alpha,
const float *  AP,
int  lda,
hipblasStride  strideA,
float *  BP,
int  ldb,
hipblasStride  strideB,
int  batchCount 
)

BLAS Level 3 API.

trsmSridedBatched performs the following strided batched operation:

op(A_i)*X_i = alpha*B_i or  X_i*op(A_i) = alpha*B_i, for i = 1, ..., batchCount.

where alpha is a scalar, X and B are strided batched m by n matrices, A is triangular strided batched matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

Each matrix X_i is overwritten on B_i for i = 1, ..., batchCount.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time. (where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each B_i. m >= 0.
[in]n[int] n specifies the number of columns of each B_i. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice pointer pointing to the first matrix A_1. of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of each A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in,out]BPdevice pointer pointing to the first matrix B_1.
[in]ldb[int] ldb specifies the first dimension of each B_i. ldb >= max( 1, m ).
[in]strideB[hipblasStride] stride from the start of one B_i matrix to the next B_(i + 1).
[in]batchCount[int] number of trsm operatons in the batch.

◆ hipblasStrsv()

hipblasStatus_t hipblasStrsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *  AP,
int  lda,
float *  x,
int  incx 
)

BLAS Level 2 API.

trsv solves

 A*x = b or A**T*x = b,

where x and b are vectors and A is a triangular matrix.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice pointer storing matrix A, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A. lda = max( 1, m ).
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasStrsvBatched()

hipblasStatus_t hipblasStrsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *const  AP[],
int  lda,
float *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

trsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i,

where (A_i, x_i, b_i) is the i-th instance of the batch. x_i and b_i are vectors and A_i is an m by m triangular matrix.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i. lda = max(1, m)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of x.
[in]batchCount[int] number of instances in the batch

◆ hipblasStrsvStridedBatched()

hipblasStatus_t hipblasStrsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const float *  AP,
int  lda,
hipblasStride  strideA,
float *  x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

trsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i,

where (A_i, x_i, b_i) is the i-th instance of the batch. x_i and b_i are vectors and A_i is an m by m triangular matrix, for i = 1, ..., batchCount.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice pointer to the first matrix (A_1) in the batch, of dimension ( lda, m )
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1)
[in]lda[int] specifies the leading dimension of each A_i. lda = max( 1, m ).
[in,out]xdevice pointer to the first vector (x_1) in the batch.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_(i + 1)
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasStrtri()

hipblasStatus_t hipblasStrtri ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const float *  AP,
int  lda,
float *  invA,
int  ldinvA 
)

BLAS Level 3 API.

trtri compute the inverse of a matrix A, namely, invA

and write the result into invA;
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int] size of matrix A and invA
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[out]invAdevice pointer storing matrix invA.
[in]ldinvA[int] specifies the leading dimension of invA.

◆ hipblasStrtriBatched()

hipblasStatus_t hipblasStrtriBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const float *const  AP[],
int  lda,
float *  invA[],
int  ldinvA,
int  batchCount 
)

BLAS Level 3 API.

trtriBatched compute the inverse of A_i and write into invA_i where A_i and invA_i are the i-th matrices in the batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER'
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int]
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[out]invAdevice array of device pointers storing the inverse of each matrix A_i. Partial inplace operation is supported, see below. If UPLO = 'U', the leading N-by-N upper triangular part of the invA will store the inverse of the upper triangular matrix, and the strictly lower triangular part of invA is cleared. If UPLO = 'L', the leading N-by-N lower triangular part of the invA will store the inverse of the lower triangular matrix, and the strictly upper triangular part of invA is cleared.
[in]ldinvA[int] specifies the leading dimension of each invA_i.
[in]batchCount[int] numbers of matrices in the batch

◆ hipblasStrtriStridedBatched()

hipblasStatus_t hipblasStrtriStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const float *  AP,
int  lda,
hipblasStride  strideA,
float *  invA,
int  ldinvA,
hipblasStride  stride_invA,
int  batchCount 
)

BLAS Level 3 API.

trtriStridedBatched compute the inverse of A_i and write into invA_i where A_i and invA_i are the i-th matrices in the batch, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER'
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int]
[in]APdevice pointer pointing to address of first matrix A_1.
[in]lda[int] specifies the leading dimension of each A.
[in]strideA[hipblasStride] "batch stride a": stride from the start of one A_i matrix to the next A_(i + 1).
[out]invAdevice pointer storing the inverses of each matrix A_i. Partial inplace operation is supported, see below. If UPLO = 'U', the leading N-by-N upper triangular part of the invA will store the inverse of the upper triangular matrix, and the strictly lower triangular part of invA is cleared. If UPLO = 'L', the leading N-by-N lower triangular part of the invA will store the inverse of the lower triangular matrix, and the strictly upper triangular part of invA is cleared.
[in]ldinvA[int] specifies the leading dimension of each invA_i.
[in]stride_invA[hipblasStride] "batch stride invA": stride from the start of one invA_i matrix to the next invA_(i + 1).
[in]batchCount[int] numbers of matrices in the batch

◆ hipblasTrsmBatchedEx()

hipblasStatus_t hipblasTrsmBatchedEx ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const void *  alpha,
void *  A,
int  lda,
void *  B,
int  ldb,
int  batchCount,
const void *  invA,
int  invAsize,
hipblasDatatype_t  computeType 
)

BLAS EX API

trsmBatchedEx solves

op(A_i)*X_i = alpha*B_i or X_i*op(A_i) = alpha*B_i,

for i = 1, ..., batchCount; and where alpha is a scalar, X and B are arrays of m by n matrices, A is an array of triangular matrix and each op(A_i) is one of

op( A_i ) = A_i   or   op( A_i ) = A_i^T   or   op( A_i ) = A_i^H.

Each matrix X_i is overwritten on B_i.

This function gives the user the ability to reuse the invA matrix between runs. If invA == NULL, hipblasTrsmBatchedEx will automatically calculate each invA_i on every run.

Setting up invA: Each accepted invA_i matrix consists of the packed 128x128 inverses of the diagonal blocks of matrix A_i, followed by any smaller diagonal block that remains. To set up each invA_i it is recommended that hipblasTrtriBatched be used with matrix A_i as the input. invA is an array of pointers of batchCount length holding each invA_i.

Device memory of size 128 x k should be allocated for each invA_i ahead of time, where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT. The actual number of elements in each invA_i should be passed as invAsize.

To begin, hipblasTrtriBatched must be called on the full 128x128 sized diagonal blocks of each matrix A_i. Below are the restricted parameters:

  • n = 128
  • ldinvA = 128
  • stride_invA = 128x128
  • batchCount = k / 128,

Then any remaining block may be added:

  • n = k % 128
  • invA = invA + stride_invA * previousBatchCount
  • ldinvA = 128
  • batchCount = 1
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each B_i. m >= 0.
[in]n[int] n specifies the number of columns of each B_i. n >= 0.
[in]alpha[void *] device pointer or host pointer alpha specifying the scalar alpha. When alpha is &zero then A is not referenced, and B need not be set before entry.
[in]A[void *] device array of device pointers storing each matrix A_i. each A_i is of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of each A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]B[void *] device array of device pointers storing each matrix B_i. each B_i is of dimension ( ldb, n ). Before entry, the leading m by n part of the array B_i must contain the right-hand side matrix B_i, and on exit is overwritten by the solution matrix X_i
[in]ldb[int] ldb specifies the first dimension of each B_i. ldb >= max( 1, m ).
[in]batchCount[int] specifies how many batches.
[in]invA[void *] device array of device pointers storing the inverse diagonal blocks of each A_i. each invA_i is of dimension ( ld_invA, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT. ld_invA must be equal to 128.
[in]invAsize[int] invAsize specifies the number of elements of device memory in each invA_i.
[in]computeType[hipblasDatatype_t] specifies the datatype of computation

◆ hipblasTrsmEx()

hipblasStatus_t hipblasTrsmEx ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const void *  alpha,
void *  A,
int  lda,
void *  B,
int  ldb,
const void *  invA,
int  invAsize,
hipblasDatatype_t  computeType 
)

BLAS EX API

trsmEx solves

op(A)*X = alpha*B or X*op(A) = alpha*B,

where alpha is a scalar, X and B are m by n matrices, A is triangular matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

The matrix X is overwritten on B.

This function gives the user the ability to reuse the invA matrix between runs. If invA == NULL, hipblasTrsmEx will automatically calculate invA on every run.

Setting up invA: The accepted invA matrix consists of the packed 128x128 inverses of the diagonal blocks of matrix A, followed by any smaller diagonal block that remains. To set up invA it is recommended that hipblasTrtriBatched be used with matrix A as the input.

Device memory of size 128 x k should be allocated for invA ahead of time, where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT. The actual number of elements in invA should be passed as invAsize.

To begin, hipblasTrtriBatched must be called on the full 128x128 sized diagonal blocks of matrix A. Below are the restricted parameters:

  • n = 128
  • ldinvA = 128
  • stride_invA = 128x128
  • batchCount = k / 128,

Then any remaining block may be added:

  • n = k % 128
  • invA = invA + stride_invA * previousBatchCount
  • ldinvA = 128
  • batchCount = 1
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_ON_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B. m >= 0.
[in]n[int] n specifies the number of columns of B. n >= 0.
[in]alpha[void *] device pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced, and B need not be set before entry.
[in]A[void *] device pointer storing matrix A. of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]B[void *] device pointer storing matrix B. B is of dimension ( ldb, n ). Before entry, the leading m by n part of the array B must contain the right-hand side matrix B, and on exit is overwritten by the solution matrix X.
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m ).
[in]invA[void *] device pointer storing the inverse diagonal blocks of A. invA is of dimension ( ld_invA, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT. ld_invA must be equal to 128.
[in]invAsize[int] invAsize specifies the number of elements of device memory in invA.
[in]computeType[hipblasDatatype_t] specifies the datatype of computation

◆ hipblasTrsmStridedBatchedEx()

hipblasStatus_t hipblasTrsmStridedBatchedEx ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const void *  alpha,
void *  A,
int  lda,
hipblasStride  strideA,
void *  B,
int  ldb,
hipblasStride  strideB,
int  batchCount,
const void *  invA,
int  invAsize,
hipblasStride  strideInvA,
hipblasDatatype_t  computeType 
)

BLAS EX API

trsmStridedBatchedEx solves

op(A_i)*X_i = alpha*B_i or X_i*op(A_i) = alpha*B_i,

for i = 1, ..., batchCount; and where alpha is a scalar, X and B are strided batched m by n matrices, A is a strided batched triangular matrix and op(A_i) is one of

op( A_i ) = A_i   or   op( A_i ) = A_i^T   or   op( A_i ) = A_i^H.

Each matrix X_i is overwritten on B_i.

This function gives the user the ability to reuse each invA_i matrix between runs. If invA == NULL, hipblasTrsmStridedBatchedEx will automatically calculate each invA_i on every run.

Setting up invA: Each accepted invA_i matrix consists of the packed 128x128 inverses of the diagonal blocks of matrix A_i, followed by any smaller diagonal block that remains. To set up invA_i it is recommended that hipblasTrtriBatched be used with matrix A_i as the input. invA is a contiguous piece of memory holding each invA_i.

Device memory of size 128 x k should be allocated for each invA_i ahead of time, where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT. The actual number of elements in each invA_i should be passed as invAsize.

To begin, hipblasTrtriBatched must be called on the full 128x128 sized diagonal blocks of each matrix A_i. Below are the restricted parameters:

  • n = 128
  • ldinvA = 128
  • stride_invA = 128x128
  • batchCount = k / 128,

Then any remaining block may be added:

  • n = k % 128
  • invA = invA + stride_invA * previousBatchCount
  • ldinvA = 128
  • batchCount = 1
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each B_i. m >= 0.
[in]n[int] n specifies the number of columns of each B_i. n >= 0.
[in]alpha[void *] device pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced, and B need not be set before entry.
[in]A[void *] device pointer storing matrix A. of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in]strideA[hipblasStride] The stride between each A matrix.
[in,out]B[void *] device pointer pointing to first matrix B_i. each B_i is of dimension ( ldb, n ). Before entry, the leading m by n part of each array B_i must contain the right-hand side of matrix B_i, and on exit is overwritten by the solution matrix X_i.
[in]ldb[int] ldb specifies the first dimension of each B_i. ldb >= max( 1, m ).
[in]strideB[hipblasStride] The stride between each B_i matrix.
[in]batchCount[int] specifies how many batches.
[in]invA[void *] device pointer storing the inverse diagonal blocks of each A_i. invA points to the first invA_1. each invA_i is of dimension ( ld_invA, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT. ld_invA must be equal to 128.
[in]invAsize[int] invAsize specifies the number of elements of device memory in each invA_i.
[in]strideInvA[hipblasStride] The stride between each invA matrix.
[in]computeType[hipblasDatatype_t] specifies the datatype of computation

◆ hipblasZaxpy()

hipblasStatus_t hipblasZaxpy ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasDoubleComplex y,
int  incy 
)

BLAS Level 1 API.

axpy computes constant alpha multiplied by vector x, plus vector y

y := alpha * x + y
  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]alphadevice pointer or host pointer to specify the scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[out]ydevice pointer storing vector y.
[in,out]incy[int] specifies the increment for the elements of y.

◆ hipblasZaxpyBatched()

hipblasStatus_t hipblasZaxpyBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  x[],
int  incx,
hipblasDoubleComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

axpyBatched compute y := alpha * x + y over a set of batched vectors.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]alphaspecifies the scalar alpha.
[in]xpointer storing vector x on the GPU.
[in]incx[int] specifies the increment for the elements of x.
[out]ypointer storing vector y on the GPU.
[in,out]incy[int] specifies the increment for the elements of y.
[in]batchCount[int] number of instances in the batch

◆ hipblasZaxpyStridedBatched()

hipblasStatus_t hipblasZaxpyStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

axpyStridedBatched compute y := alpha * x + y over a set of strided batched vectors.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int]
[in]alphaspecifies the scalar alpha.
[in]xpointer storing vector x on the GPU.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] specifies the increment between vectors of x.
[out]ypointer storing vector y on the GPU.
[in,out]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] specifies the increment between vectors of y.
[in]batchCount[int] number of instances in the batch

◆ hipblasZcopy()

hipblasStatus_t hipblasZcopy ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
hipblasDoubleComplex y,
int  incy 
)

BLAS Level 1 API.

copy copies each element x[i] into y[i], for i = 1 , ... , n

y := x,
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x to be copied to y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasZcopyBatched()

hipblasStatus_t hipblasZcopyBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex *const  x[],
int  incx,
hipblasDoubleComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

copyBatched copies each element x_i[j] into y_i[j], for j = 1 , ... , n; i = 1 , ... , batchCount

y_i := x_i,

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i to be copied to y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasZcopyStridedBatched()

hipblasStatus_t hipblasZcopyStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

copyStridedBatched copies each element x_i[j] into y_i[j], for j = 1 , ... , n; i = 1 , ... , batchCount

y_i := x_i,

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i to be copied to y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increments for the elements of vectors x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of vectors y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stride_y, however the user should take care to ensure that stride_y is of appropriate size, for a typical case this means stride_y >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasZdgmm()

hipblasStatus_t hipblasZdgmm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex x,
int  incx,
hipblasDoubleComplex CP,
int  ldc 
)

BLAS Level 3 API.

dgmm performs one of the matrix-matrix operations

C = A * diag(x) if side == HIPBLAS_SIDE_RIGHT
C = diag(x) * A if side == HIPBLAS_SIDE_LEFT

where C and A are m by n dimensional matrices. diag( x ) is a diagonal matrix and x is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between values of x
[in,out]CPdevice pointer storing matrix C.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasZdgmmBatched()

hipblasStatus_t hipblasZdgmmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  x[],
int  incx,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

dgmmBatched performs one of the batched matrix-matrix operations

C_i = A_i * diag(x_i) for i = 0, 1, ... batchCount-1 if side == HIPBLAS_SIDE_RIGHT
C_i = diag(x_i) * A_i for i = 0, 1, ... batchCount-1 if side == HIPBLAS_SIDE_LEFT

where C_i and A_i are m by n dimensional matrices. diag(x_i) is a diagonal matrix and x_i is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, n )
[in]lda[int] specifies the leading dimension of A_i.
[in]xdevice array of device pointers storing each vector x_i on the GPU. Each x_i is of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT
[in]incx[int] specifies the increment between values of x_i
[in,out]CPdevice array of device pointers storing each matrix C_i on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasZdgmmStridedBatched()

hipblasStatus_t hipblasZdgmmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
int  m,
int  n,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
hipblasDoubleComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

dgmmStridedBatched performs one of the batched matrix-matrix operations

C_i = A_i * diag(x_i)   if side == HIPBLAS_SIDE_RIGHT   for i = 0, 1, ... batchCount-1
C_i = diag(x_i) * A_i   if side == HIPBLAS_SIDE_LEFT    for i = 0, 1, ... batchCount-1

where C_i and A_i are m by n dimensional matrices. diag(x_i) is a diagonal matrix and x_i is vector of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] specifies the side of diag(x)
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]APdevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, n )
[in]lda[int] specifies the leading dimension of A.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xpointer to the first vector x_0 on the GPU. Each x_i is of dimension n if side == HIPBLAS_SIDE_RIGHT and dimension m if side == HIPBLAS_SIDE_LEFT
[in]incx[int] specifies the increment between values of x
[in]stridex[hipblasStride] stride from the start of one vector(x_i) and the next one (x_i+1)
[in,out]CPdevice pointer to the first matrix C_0 on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasZdotc()

hipblasStatus_t hipblasZdotc ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex y,
int  incy,
hipblasDoubleComplex result 
)

BLAS Level 1 API.

dot(u) performs the dot product of vectors x and y

result = x * y;

dotc performs the dot product of the conjugate of complex vector x and complex vector y

result = conjugate (x) * y;
  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the dot product. return is 0.0 if n <= 0.

◆ hipblasZdotcBatched()

hipblasStatus_t hipblasZdotcBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex *const  y[],
int  incy,
int  batchCount,
hipblasDoubleComplex result 
)

BLAS Level 1 API.

dotBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasZdotcStridedBatched()

hipblasStatus_t hipblasZdotcStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount,
hipblasDoubleComplex result 
)

BLAS Level 1 API.

dotStridedBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcStridedBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasZdotu()

hipblasStatus_t hipblasZdotu ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex y,
int  incy,
hipblasDoubleComplex result 
)

BLAS Level 1 API.

dot(u) performs the dot product of vectors x and y

result = x * y;

dotc performs the dot product of the conjugate of complex vector x and complex vector y

result = conjugate (x) * y;
  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of y.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]resultdevice pointer or host pointer to store the dot product. return is 0.0 if n <= 0.

◆ hipblasZdotuBatched()

hipblasStatus_t hipblasZdotuBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex *const  y[],
int  incy,
int  batchCount,
hipblasDoubleComplex result 
)

BLAS Level 1 API.

dotBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasZdotuStridedBatched()

hipblasStatus_t hipblasZdotuStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount,
hipblasDoubleComplex result 
)

BLAS Level 1 API.

dotStridedBatched(u) performs a batch of dot products of vectors x and y

result_i = x_i * y_i;

dotcStridedBatched performs a batch of dot products of the conjugate of complex vector x and complex vector y

result_i = conjugate (x_i) * y_i;

where (x_i, y_i) is the i-th instance of the batch. x_i and y_i are vectors, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : h,bf,s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch
[in,out]resultdevice array or host array of batchCount size to store the dot products of each batch. return 0.0 for each element if n <= 0.

◆ hipblasZdrot()

hipblasStatus_t hipblasZdrot ( hipblasHandle_t  handle,
int  n,
hipblasDoubleComplex x,
int  incx,
hipblasDoubleComplex y,
int  incy,
const double *  c,
const double *  s 
)

BLAS Level 1 API.

rot applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to vectors x and y. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment between elements of y.
[in]cdevice pointer or host pointer storing scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer storing scalar sine component of the rotation matrix.

◆ hipblasZdrotBatched()

hipblasStatus_t hipblasZdrotBatched ( hipblasHandle_t  handle,
int  n,
hipblasDoubleComplex *const  x[],
int  incx,
hipblasDoubleComplex *const  y[],
int  incy,
const double *  c,
const double *  s,
int  batchCount 
)

BLAS Level 1 API.

rotBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice array of deivce pointers storing each vector x_i.
[in]incx[int] specifies the increment between elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment between elements of each y_i.
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasZdrotStridedBatched()

hipblasStatus_t hipblasZdrotStridedBatched ( hipblasHandle_t  handle,
int  n,
hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
const double *  c,
const double *  s,
int  batchCount 
)

BLAS Level 1 API.

rotStridedBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to strided batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment between elements of each x_i.
[in]stridex[hipblasStride] specifies the increment from the beginning of x_i to the beginning of x_(i+1)
[in,out]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment between elements of each y_i.
[in]stridey[hipblasStride] specifies the increment from the beginning of y_i to the beginning of y_(i+1)
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasZdscal()

hipblasStatus_t hipblasZdscal ( hipblasHandle_t  handle,
int  n,
const double *  alpha,
hipblasDoubleComplex x,
int  incx 
)

BLAS Level 1 API.

scal scales each element of vector x with scalar alpha.

x := alpha * x
  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]alphadevice pointer or host pointer for the scalar alpha.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasZdscalBatched()

hipblasStatus_t hipblasZdscalBatched ( hipblasHandle_t  handle,
int  n,
const double *  alpha,
hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 1 API.

scalBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasZdscalStridedBatched()

hipblasStatus_t hipblasZdscalStridedBatched ( hipblasHandle_t  handle,
int  n,
const double *  alpha,
hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 1 API.

scalStridedBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i ,

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasZgbmv()

hipblasStatus_t hipblasZgbmv ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy 
)

BLAS Level 2 API.

gbmv performs one of the matrix-vector operations

y := alpha*A*x    + beta*y,   or
y := alpha*A**T*x + beta*y,   or
y := alpha*A**H*x + beta*y,

where alpha and beta are scalars, x and y are vectors and A is an m by n banded matrix with kl sub-diagonals and ku super-diagonals.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]kl[int] number of sub-diagonals of A
[in]ku[int] number of super-diagonals of A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing banded matrix A. Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of A. Must be >= (kl + ku + 1)
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasZgbmvBatched()

hipblasStatus_t hipblasZgbmvBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

gbmvBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n banded matrix with kl sub-diagonals and ku super-diagonals, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of each matrix A_i
[in]n[int] number of columns of each matrix A_i
[in]kl[int] number of sub-diagonals of each A_i
[in]ku[int] number of super-diagonals of each A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each banded matrix A_i. Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of each A_i. Must be >= (kl + ku + 1)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasZgbmvStridedBatched()

hipblasStatus_t hipblasZgbmvStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
int  kl,
int  ku,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

gbmvStridedBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n banded matrix with kl sub-diagonals and ku super-diagonals, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]kl[int] number of sub-diagonals of A
[in]ku[int] number of super-diagonals of A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer to first banded matrix (A_1). Leading (kl + ku + 1) by n part of the matrix contains the coefficients of the banded matrix. The leading diagonal resides in row (ku + 1) with the first super-diagonal above on the RHS of row ku. The first sub-diagonal resides below on the LHS of row ku + 2. This propogates up and down across sub/super-diagonals. Ex: (m = n = 7; ku = 2, kl = 2) 1 2 3 0 0 0 0 0 0 3 3 3 3 3 4 1 2 3 0 0 0 0 2 2 2 2 2 2 5 4 1 2 3 0 0 -—> 1 1 1 1 1 1 1 0 5 4 1 2 3 0 4 4 4 4 4 4 0 0 0 5 4 1 2 0 5 5 5 5 5 0 0 0 0 0 5 4 1 2 0 0 0 0 0 0 0 0 0 0 0 5 4 1 0 0 0 0 0 0 0 Note that the empty elements which don't correspond to data will not be referenced.
[in]lda[int] specifies the leading dimension of A. Must be >= (kl + ku + 1)
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xdevice pointer to first vector (x_1).
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer to first vector (y_1).
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (x_i+1)
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasZgeam()

hipblasStatus_t hipblasZgeam ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex beta,
const hipblasDoubleComplex BP,
int  ldb,
hipblasDoubleComplex CP,
int  ldc 
)

BLAS Level 3 API.

geam performs one of the matrix-matrix operations

C = alpha*op( A ) + beta*op( B ),

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are matrices, with op( A ) an m by n matrix, op( B ) an m by n matrix, and C an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPdevice pointer storing matrix B.
[in]ldb[int] specifies the leading dimension of B.
[in,out]CPdevice pointer storing matrix C.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasZgeamBatched()

hipblasStatus_t hipblasZgeamBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex beta,
const hipblasDoubleComplex *const  BP[],
int  ldb,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

geamBatched performs one of the batched matrix-matrix operations

C_i = alpha*op( A_i ) + beta*op( B_i )  for i = 0, 1, ... batchCount - 1

where alpha and beta are scalars, and op(A_i), op(B_i) and C_i are m by n matrices and op( X ) is one of

op( X ) = X      or
op( X ) = X**T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, k ), where k is m when transA == HIPBLAS_OP_N and is n when transA == HIPBLAS_OP_T.
[in]lda[int] specifies the leading dimension of A.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPdevice array of device pointers storing each matrix B_i on the GPU. Each B_i is of dimension ( ldb, k ), where k is m when transB == HIPBLAS_OP_N and is n when transB == HIPBLAS_OP_T.
[in]ldb[int] specifies the leading dimension of B.
[in,out]CPdevice array of device pointers storing each matrix C_i on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]batchCount[int] number of instances i in the batch.

◆ hipblasZgeamStridedBatched()

hipblasStatus_t hipblasZgeamStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex beta,
const hipblasDoubleComplex BP,
int  ldb,
hipblasStride  strideB,
hipblasDoubleComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

geamStridedBatched performs one of the batched matrix-matrix operations

C_i = alpha*op( A_i ) + beta*op( B_i )  for i = 0, 1, ... batchCount - 1

where alpha and beta are scalars, and op(A_i), op(B_i) and C_i are m by n matrices and op( X ) is one of

op( X ) = X      or
op( X ) = X**T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimension m.
[in]n[int] matrix dimension n.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, k ), where k is m when transA == HIPBLAS_OP_N and is n when transA == HIPBLAS_OP_T.
[in]lda[int] specifies the leading dimension of A.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]betadevice pointer or host pointer specifying the scalar beta.
[in]BPpointer to the first matrix B_0 on the GPU. Each B_i is of dimension ( ldb, k ), where k is m when transB == HIPBLAS_OP_N and is n when transB == HIPBLAS_OP_T.
[in]ldb[int] specifies the leading dimension of B.
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in,out]CPpointer to the first matrix C_0 on the GPU. Each C_i is of dimension ( ldc, n ).
[in]ldc[int] specifies the leading dimension of C.
[in]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasZgels()

hipblasStatus_t hipblasZgels ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
hipblasDoubleComplex A,
const int  lda,
hipblasDoubleComplex B,
const int  ldb,
int *  info,
int *  deviceInfo 
)

GELS solves an overdetermined (or underdetermined) linear system defined by an m-by-n matrix A, and a corresponding matrix B, using the QR factorization computed by GEQRF (or the LQ factorization computed by "GELQF").

Depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A X = B & \: \text{not transposed, or}\\ A' X = B & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X is found by minimizing

\[ || B - A X || \quad \text{(or} \: || B - A' X ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X is chosen such that \(|| X ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : currently unsupported
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of matrix A.
[in]nint. n >= 0.
The number of columns of matrix A.
[in]nrhsint. nrhs >= 0.
The number of columns of matrices B and X; i.e., the columns on the right hand side.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the matrix A. On exit, the QR (or LQ) factorization of A as returned by "GEQRF" (or "GELQF").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrix A.
[in,out]Bpointer to type. Array on the GPU of dimension ldb*nrhs.
On entry, the matrix B. On exit, when info = 0, B is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrix B.
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int on the GPU.
If info = 0, successful exit. If info = i > 0, the solution could not be computed because input matrix A is rank deficient; the i-th diagonal element of its triangular factor is zero.

◆ hipblasZgelsBatched()

hipblasStatus_t hipblasZgelsBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
hipblasDoubleComplex *const  A[],
const int  lda,
hipblasDoubleComplex *const  B[],
const int  ldb,
int *  info,
int *  deviceInfo,
const int  batchCount 
)

gelsBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_BATCHED" (or the LQ factorizations computed by "GELQF_BATCHED").

For each instance in the batch, depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A_j X_j = B_j & \: \text{not transposed, or}\\ A_j' X_j = B_j & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X_j is found by minimizing

\[ || B_j - A_j X_j || \quad \text{(or} \: || B_j - A_j' X_j ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X_j is chosen such that \(|| X_j ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z Note that cuBLAS backend supports only the non-transpose operation and only solves over-determined systems (m >= n).
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of all matrices A_j in the batch.
[in]nint. n >= 0.
The number of columns of all matrices A_j in the batch.
[in]nrhsint. nrhs >= 0.
The number of columns of all matrices B_j and X_j in the batch; i.e., the columns on the right hand side.
[in,out]Aarray of pointer to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the matrices A_j. On exit, the QR (or LQ) factorizations of A_j as returned by "GEQRF_BATCHED" (or "GELQF_BATCHED").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_j.
[in,out]Barray of pointer to type. Each pointer points to an array on the GPU of dimension ldb*nrhs.
On entry, the matrices B_j. On exit, when info[j] = 0, B_j is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrices B_j.
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int. Array of batchCount integers on the GPU.
If deviceInfo[j] = 0, successful exit for solution of A_j. If deviceInfo[j] = i > 0, the solution of A_j could not be computed because input matrix A_j is rank deficient; the i-th diagonal element of its triangular factor is zero.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasZgelsStridedBatched()

hipblasStatus_t hipblasZgelsStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
const int  m,
const int  n,
const int  nrhs,
hipblasDoubleComplex A,
const int  lda,
const hipblasStride  strideA,
hipblasDoubleComplex B,
const int  ldb,
const hipblasStride  strideB,
int *  info,
int *  deviceInfo,
const int  batch_count 
)

gelsStridedBatched solves a batch of overdetermined (or underdetermined) linear systems defined by a set of m-by-n matrices \(A_j\), and corresponding matrices \(B_j\), using the QR factorizations computed by "GEQRF_STRIDED_BATCHED" (or the LQ factorizations computed by "GELQF_STRIDED_BATCHED").

For each instance in the batch, depending on the value of trans, the problem solved by this function is either of the form

\[ \begin{array}{cl} A_j X_j = B_j & \: \text{not transposed, or}\\ A_j' X_j = B_j & \: \text{transposed if real, or conjugate transposed if complex} \end{array} \]

If m >= n (or m < n in the case of transpose/conjugate transpose), the system is overdetermined and a least-squares solution approximating X_j is found by minimizing

\[ || B_j - A_j X_j || \quad \text{(or} \: || B_j - A_j' X_j ||\text{)} \]

If m < n (or m >= n in the case of transpose/conjugate transpose), the system is underdetermined and a unique solution for X_j is chosen such that \(|| X_j ||\) is minimal.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : currently unsupported
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]mint. m >= 0.
The number of rows of all matrices A_j in the batch.
[in]nint. n >= 0.
The number of columns of all matrices A_j in the batch.
[in]nrhsint. nrhs >= 0.
The number of columns of all matrices B_j and X_j in the batch; i.e., the columns on the right hand side.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the matrices A_j. On exit, the QR (or LQ) factorizations of A_j as returned by "GEQRF_STRIDED_BATCHED" (or "GELQF_STRIDED_BATCHED").
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_j.
[in]strideAhipblasStride.
Stride from the start of one matrix A_j to the next one A_(j+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n
[in,out]Bpointer to type. Array on the GPU (the size depends on the value of strideB).
On entry, the matrices B_j. On exit, when info[j] = 0, each B_j is overwritten by the solution vectors (and the residuals in the overdetermined cases) stored as columns.
[in]ldbint. ldb >= max(m,n).
Specifies the leading dimension of matrices B_j.
[in]strideBhipblasStride.
Stride from the start of one matrix B_j to the next one B_(j+1). There is no restriction for the value of strideB. Normal use case is strideB >= ldb*nrhs
[out]infopointer to an int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[out]deviceInfopointer to int. Array of batchCount integers on the GPU.
If deviceInfo[j] = 0, successful exit for solution of A_j. If deviceInfo[j] = i > 0, the solution of A_j could not be computed because input matrix A_j is rank deficient; the i-th diagonal element of its triangular factor is zero.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasZgemm()

hipblasStatus_t hipblasZgemm ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex BP,
int  ldb,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc 
)

BLAS Level 3 API.

gemm performs one of the matrix-matrix operations

C = alpha*op( A )*op( B ) + beta*C,

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are matrices, with op( A ) an m by k matrix, op( B ) a k by n matrix and C an m by n matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t]
      .
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] number or rows of matrices op( A ) and C
[in]n[int] number of columns of matrices op( B ) and C
[in]k[int] number of columns of matrix op( A ) and number of rows of matrix op( B )
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]BPdevice pointer storing matrix B.
[in]ldb[int] specifies the leading dimension of B.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice pointer storing matrix C on the GPU.
[in]ldc[int] specifies the leading dimension of C.

◆ hipblasZgemmBatched()

hipblasStatus_t hipblasZgemmBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  BP[],
int  ldb,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

gemmBatched performs one of the batched matrix-matrix operations C_i = alpha*op( A_i )*op( B_i ) + beta*C_i, for i = 1, ..., batchCount. where op( X ) is one of op( X ) = X or op( X ) = X**T or op( X ) = X**H, alpha and beta are scalars, and A, B and C are strided batched matrices, with op( A ) an m by k by batchCount strided_batched matrix, op( B ) an k by n by batchCount strided_batched matrix and C an m by n by batchCount strided_batched matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimention m.
[in]n[int] matrix dimention n.
[in]k[int] matrix dimention k.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]BPdevice array of device pointers storing each matrix B_i.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice array of device pointers storing each matrix C_i.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]batchCount[int] number of gemm operations in the batch

◆ hipblasZgemmStridedBatched()

hipblasStatus_t hipblasZgemmStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
hipblasOperation_t  transB,
int  m,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
long long  strideA,
const hipblasDoubleComplex BP,
int  ldb,
long long  strideB,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc,
long long  strideC,
int  batchCount 
)

BLAS Level 3 API.

gemmStridedBatched performs one of the strided batched matrix-matrix operations

C_i = alpha*op( A_i )*op( B_i ) + beta*C_i, for i = 1, ..., batchCount.

where op( X ) is one of

op( X ) = X      or
op( X ) = X**T   or
op( X ) = X**H,

alpha and beta are scalars, and A, B and C are strided batched matrices, with op( A ) an m by k by batchCount strided_batched matrix, op( B ) an k by n by batchCount strided_batched matrix and C an m by n by batchCount strided_batched matrix.

  • Supported precisions in rocBLAS : h,s,d,c,z
  • Supported precisions in cuBLAS : h,s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] specifies the form of op( A )
[in]transB[hipblasOperation_t] specifies the form of op( B )
[in]m[int] matrix dimention m.
[in]n[int] matrix dimention n.
[in]k[int] matrix dimention k.
[in]alphadevice pointer or host pointer specifying the scalar alpha.
[in]APdevice pointer pointing to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in]BPdevice pointer pointing to the first matrix B_1.
[in]ldb[int] specifies the leading dimension of each B_i.
[in]strideB[hipblasStride] stride from the start of one B_i matrix to the next B_(i + 1).
[in]betadevice pointer or host pointer specifying the scalar beta.
[in,out]CPdevice pointer pointing to the first matrix C_1.
[in]ldc[int] specifies the leading dimension of each C_i.
[in]strideC[hipblasStride] stride from the start of one C_i matrix to the next C_(i + 1).
[in]batchCount[int] number of gemm operatons in the batch

◆ hipblasZgemv()

hipblasStatus_t hipblasZgemv ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy 
)

BLAS Level 2 API.

gemv performs one of the matrix-vector operations

y := alpha*A*x    + beta*y,   or
y := alpha*A**T*x + beta*y,   or
y := alpha*A**H*x + beta*y,

where alpha and beta are scalars, x and y are vectors and A is an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not
[in]m[int] number of rows of matrix A
[in]n[int] number of columns of matrix A
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasZgemvBatched()

hipblasStatus_t hipblasZgemvBatched ( hipblasHandle_t  handle,
hipblasOperation_t  trans,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

gemvBatched performs a batch of matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]trans[hipblasOperation_t] indicates whether matrices A_i are tranposed (conjugated) or not
[in]m[int] number of rows of each matrix A_i
[in]n[int] number of columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each matrix A_i.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasZgemvStridedBatched()

hipblasStatus_t hipblasZgemvStridedBatched ( hipblasHandle_t  handle,
hipblasOperation_t  transA,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

gemvStridedBatched performs a batch of matrix-vector operations

y_i := alpha*A_i*x_i    + beta*y_i,   or
y_i := alpha*A_i**T*x_i + beta*y_i,   or
y_i := alpha*A_i**H*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]transA[hipblasOperation_t] indicates whether matrices A_i are tranposed (conjugated) or not
[in]m[int] number of rows of matrices A_i
[in]n[int] number of columns of matrices A_i
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer to the first matrix (A_1) in the batch.
[in]lda[int] specifies the leading dimension of matrices A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. When trans equals HIPBLAS_OP_N this typically means stridex >= n * incx, otherwise stridex >= m * incx.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of vectors y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. When trans equals HIPBLAS_OP_N this typically means stridey >= m * incy, otherwise stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasZgeqrf()

hipblasStatus_t hipblasZgeqrf ( hipblasHandle_t  handle,
const int  m,
const int  n,
hipblasDoubleComplex A,
const int  lda,
hipblasDoubleComplex ipiv,
int *  info 
)

SOLVER API.

geqrf computes a QR factorization of a general m-by-n matrix A.

The factorization has the form

\[ A = Q\left[\begin{array}{c} R\\ 0 \end{array}\right] \]

where R is upper triangular (upper trapezoidal if m < n), and Q is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q = H_1H_2\cdots H_k, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_i\) is given by

\[ H_i = I - \text{ipiv}[i] \cdot v_i v_i' \]

where the first i-1 elements of the Householder vector \(v_i\) are zero, and \(v_i[i] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of the matrix A.
[in]nint. n >= 0.
The number of columns of the matrix A.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the m-by-n matrix to be factored. On exit, the elements on and above the diagonal contain the factor R; the elements below the diagonal are the last m - i elements of Householder vector v_i.
[in]ldaint. lda >= m.
Specifies the leading dimension of A.
[out]ipivpointer to type. Array on the GPU of dimension min(m,n).
The Householder scalars.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.

◆ hipblasZgeqrfBatched()

hipblasStatus_t hipblasZgeqrfBatched ( hipblasHandle_t  handle,
const int  m,
const int  n,
hipblasDoubleComplex *const  A[],
const int  lda,
hipblasDoubleComplex *const  ipiv[],
int *  info,
const int  batchCount 
)

SOLVER API.

geqrfBatched computes the QR factorization of a batch of general m-by-n matrices.

The factorization of matrix \(A_i\) in the batch has the form

\[ A_i = Q_i\left[\begin{array}{c} R_i\\ 0 \end{array}\right] \]

where \(R_i\) is upper triangular (upper trapezoidal if m < n), and \(Q_i\) is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q_i = H_{i_1}H_{i_2}\cdots H_{i_k}, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_{i_j}\) is given by

\[ H_{i_j} = I - \text{ipiv}_i[j] \cdot v_{i_j} v_{i_j}' \]

where the first j-1 elements of Householder vector \(v_{i_j}\) are zero, and \(v_{i_j}[j] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of all the matrices A_i in the batch.
[in]nint. n >= 0.
The number of columns of all the matrices A_i in the batch.
[in,out]AArray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the m-by-n matrices A_i to be factored. On exit, the elements on and above the diagonal contain the factor R_i. The elements below the diagonal are the last m - j elements of Householder vector v_(i_j).
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_i.
[out]ipivarray of pointers to type. Each pointer points to an array on the GPU of dimension min(m, n).
Contains the vectors ipiv_i of corresponding Householder scalars.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasZgeqrfStridedBatched()

hipblasStatus_t hipblasZgeqrfStridedBatched ( hipblasHandle_t  handle,
const int  m,
const int  n,
hipblasDoubleComplex A,
const int  lda,
const hipblasStride  strideA,
hipblasDoubleComplex ipiv,
const hipblasStride  strideP,
int *  info,
const int  batchCount 
)

SOLVER API.

geqrfStridedBatched computes the QR factorization of a batch of general m-by-n matrices.

The factorization of matrix \(A_i\) in the batch has the form

\[ A_i = Q_i\left[\begin{array}{c} R_i\\ 0 \end{array}\right] \]

where \(R_i\) is upper triangular (upper trapezoidal if m < n), and \(Q_i\) is a m-by-m orthogonal/unitary matrix represented as the product of Householder matrices

\[ Q_i = H_{i_1}H_{i_2}\cdots H_{i_k}, \quad \text{with} \: k = \text{min}(m,n) \]

Each Householder matrix \(H_{i_j}\) is given by

\[ H_{i_j} = I - \text{ipiv}_j[j] \cdot v_{i_j} v_{i_j}' \]

where the first j-1 elements of Householder vector \(v_{i_j}\) are zero, and \(v_{i_j}[j] = 1\).

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handlehipblasHandle_t.
[in]mint. m >= 0.
The number of rows of all the matrices A_i in the batch.
[in]nint. n >= 0.
The number of columns of all the matrices A_i in the batch.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the m-by-n matrices A_i to be factored. On exit, the elements on and above the diagonal contain the factor R_i. The elements below the diagonal are the last m - j elements of Householder vector v_(i_j).
[in]ldaint. lda >= m.
Specifies the leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n.
[out]ipivpointer to type. Array on the GPU (the size depends on the value of strideP).
Contains the vectors ipiv_i of corresponding Householder scalars.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use is strideP >= min(m,n).
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasZgerc()

hipblasStatus_t hipblasZgerc ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex y,
int  incy,
hipblasDoubleComplex AP,
int  lda 
)

BLAS Level 2 API.

ger,geru,gerc performs the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where alpha is a scalar, x and y are vectors, and A is an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of the matrix A.
[in]n[int] the number of columns of the matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasZgercBatched()

hipblasStatus_t hipblasZgercBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex *const  y[],
int  incy,
hipblasDoubleComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

gerBatched,geruBatched,gercBatched performs a batch of the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of eaceh matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasZgercStridedBatched()

hipblasStatus_t hipblasZgercStridedBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

gerStridedBatched,geruStridedBatched,gercStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**T, OR
A_i := A_i + alpha*x_i*y_i**H  for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increments for the elements of each vector x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size, for a typical case this means stridex >= m * incx.
[in,out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size, for a typical case this means stridey >= n * incy.
[in,out]APdevice pointer to the first matrix (A_1) in the batch.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasZgeru()

hipblasStatus_t hipblasZgeru ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex y,
int  incy,
hipblasDoubleComplex AP,
int  lda 
)

BLAS Level 2 API.

ger,geru,gerc performs the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where alpha is a scalar, x and y are vectors, and A is an m by n matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of the matrix A.
[in]n[int] the number of columns of the matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasZgeruBatched()

hipblasStatus_t hipblasZgeruBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex *const  y[],
int  incy,
hipblasDoubleComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

gerBatched,geruBatched,gercBatched performs a batch of the matrix-vector operations

A := A + alpha*x*y**T , OR
A := A + alpha*x*y**H for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of eaceh matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each vector x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasZgeruStridedBatched()

hipblasStatus_t hipblasZgeruStridedBatched ( hipblasHandle_t  handle,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

gerStridedBatched,geruStridedBatched,gercStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**T, OR
A_i := A_i + alpha*x_i*y_i**H  for gerc

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha is a scalar, x_i and y_i are vectors and A_i is an m by n matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]m[int] the number of rows of each matrix A_i.
[in]n[int] the number of columns of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increments for the elements of each vector x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size, for a typical case this means stridex >= m * incx.
[in,out]ydevice pointer to the first vector (y_1) in the batch.
[in]incy[int] specifies the increment for the elements of each vector y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size, for a typical case this means stridey >= n * incy.
[in,out]APdevice pointer to the first matrix (A_1) in the batch.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasZgetrf()

hipblasStatus_t hipblasZgetrf ( hipblasHandle_t  handle,
const int  n,
hipblasDoubleComplex A,
const int  lda,
int *  ipiv,
int *  info 
)

SOLVER API.

getrf computes the LU factorization of a general n-by-n matrix A using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization has the form:

\[ A = PLU \]

where P is a permutation matrix, L is lower triangular with unit diagonal elements, and U is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A = LU \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of the matrix A.
[in,out]Apointer to type. Array on the GPU of dimension lda*n.
On entry, the n-by-n matrix A to be factored. On exit, the factors L and U from the factorization. The unit diagonal elements of L are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of A.
[out]ipivpointer to int. Array on the GPU of dimension n.
The vector of pivot indices. Elements of ipiv are 1-based indices. For 1 <= i <= n, the row i of the matrix was interchanged with row ipiv[i]. Matrix P of the factorization can be derived from ipiv. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[out]infopointer to a int on the GPU.
If info = 0, successful exit. If info = j > 0, U is singular. U[j,j] is the first zero pivot.

◆ hipblasZgetrfBatched()

hipblasStatus_t hipblasZgetrfBatched ( hipblasHandle_t  handle,
const int  n,
hipblasDoubleComplex *const  A[],
const int  lda,
int *  ipiv,
int *  info,
const int  batchCount 
)

SOLVER API.

getrfBatched computes the LU factorization of a batch of general n-by-n matrices using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization of matrix \(A_i\) in the batch has the form:

\[ A_i = P_iL_iU_i \]

where \(P_i\) is a permutation matrix, \(L_i\) is lower triangular with unit diagonal elements, and \(U_i\) is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A_i = L_iU_i \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of all matrices A_i in the batch.
[in,out]Aarray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
On entry, the n-by-n matrices A_i to be factored. On exit, the factors L_i and U_i from the factorizations. The unit diagonal elements of L_i are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[out]ipivpointer to int. Array on the GPU.
Contains the vectors of pivot indices ipiv_i (corresponding to A_i). Dimension of ipiv_i is n. Elements of ipiv_i are 1-based indices. For each instance A_i in the batch and for 1 <= j <= n, the row j of the matrix A_i was interchanged with row ipiv_i[j]. Matrix P_i of the factorization can be derived from ipiv_i. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for factorization of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasZgetrfStridedBatched()

hipblasStatus_t hipblasZgetrfStridedBatched ( hipblasHandle_t  handle,
const int  n,
hipblasDoubleComplex A,
const int  lda,
const hipblasStride  strideA,
int *  ipiv,
const hipblasStride  strideP,
int *  info,
const int  batchCount 
)

SOLVER API.

getrfStridedBatched computes the LU factorization of a batch of general n-by-n matrices using partial pivoting with row interchanges. The LU factorization can be done without pivoting if ipiv is passed as a nullptr.

In the case that ipiv is not null, the factorization of matrix \(A_i\) in the batch has the form:

\[ A_i = P_iL_iU_i \]

where \(P_i\) is a permutation matrix, \(L_i\) is lower triangular with unit diagonal elements, and \(U_i\) is upper triangular.

In the case that ipiv is null, the factorization is done without pivoting:

\[ A_i = L_iU_i \]

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of columns and rows of all matrices A_i in the batch.
[in,out]Apointer to type. Array on the GPU (the size depends on the value of strideA).
On entry, the n-by-n matrices A_i to be factored. On exit, the factors L_i and U_i from the factorization. The unit diagonal elements of L_i are not stored.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n
[out]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
Contains the vectors of pivots indices ipiv_i (corresponding to A_i). Dimension of ipiv_i is n. Elements of ipiv_i are 1-based indices. For each instance A_i in the batch and for 1 <= j <= n, the row j of the matrix A_i was interchanged with row ipiv_i[j]. Matrix P_i of the factorization can be derived from ipiv_i. The factorization here can be done without pivoting if ipiv is passed in as a nullptr.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use case is strideP >= n.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for factorization of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasZgetriBatched()

hipblasStatus_t hipblasZgetriBatched ( hipblasHandle_t  handle,
const int  n,
hipblasDoubleComplex *const  A[],
const int  lda,
int *  ipiv,
hipblasDoubleComplex *const  C[],
const int  ldc,
int *  info,
const int  batchCount 
)

SOLVER API.

getriBatched computes the inverse \(C_i = A_i^{-1}\) of a batch of general n-by-n matrices \(A_i\).

The inverse is computed by solving the linear system

\[ A_i C_i = I \]

where I is the identity matrix, and \(A_i\) is factorized as \(A_i = P_i L_i U_i\) as given by getrfBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]nint. n >= 0.
The number of rows and columns of all matrices A_i in the batch.
[in]Aarray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfBatched.
[in]ldaint. lda >= n.
Specifies the leading dimension of matrices A_i.
[in]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
The pivot indices returned by getrfBatched. ipiv can be passed in as a nullptr, this will assume that getrfBatched was called without partial pivoting.
[out]Carray of pointers to type. Each pointer points to an array on the GPU of dimension ldc*n.
If info[i] = 0, the inverse of matrices A_i. Otherwise, undefined.
[in]ldcint. ldc >= n.
Specifies the leading dimension of C_i.
[out]infopointer to int. Array of batchCount integers on the GPU.
If info[i] = 0, successful exit for inversion of A_i. If info[i] = j > 0, U_i is singular. U_i[j,j] is the first zero pivot.
[in]batchCountint. batchCount >= 0.
Number of matrices in the batch.

◆ hipblasZgetrs()

hipblasStatus_t hipblasZgetrs ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
hipblasDoubleComplex A,
const int  lda,
const int *  ipiv,
hipblasDoubleComplex B,
const int  ldb,
int *  info 
)

SOLVER API.

getrs solves a system of n linear equations on n variables in its factorized form.

It solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A X = B & \: \text{not transposed,}\\ A^T X = B & \: \text{transposed, or}\\ A^H X = B & \: \text{conjugate transposed.} \end{array} \]

Matrix A is defined by its triangular factors as returned by getrf.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of A.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of the matrix B.
[in]Apointer to type. Array on the GPU of dimension lda*n.
The factors L and U of the factorization A = P*L*U returned by getrf.
[in]ldaint. lda >= n.
The leading dimension of A.
[in]ipivpointer to int. Array on the GPU of dimension n.
The pivot indices returned by getrf.
[in,out]Bpointer to type. Array on the GPU of dimension ldb*nrhs.
On entry, the right hand side matrix B. On exit, the solution matrix X.
[in]ldbint. ldb >= n.
The leading dimension of B.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.

◆ hipblasZgetrsBatched()

hipblasStatus_t hipblasZgetrsBatched ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
hipblasDoubleComplex *const  A[],
const int  lda,
const int *  ipiv,
hipblasDoubleComplex *const  B[],
const int  ldb,
int *  info,
const int  batchCount 
)

SOLVER API.

getrsBatched solves a batch of systems of n linear equations on n variables in its factorized forms.

For each instance i in the batch, it solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A_i X_i = B_i & \: \text{not transposed,}\\ A_i^T X_i = B_i & \: \text{transposed, or}\\ A_i^H X_i = B_i & \: \text{conjugate transposed.} \end{array} \]

Matrix \(A_i\) is defined by its triangular factors as returned by getrfBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations of each instance in the batch.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of all A_i matrices.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of all the matrices B_i.
[in]AArray of pointers to type. Each pointer points to an array on the GPU of dimension lda*n.
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfBatched.
[in]ldaint. lda >= n.
The leading dimension of matrices A_i.
[in]ipivpointer to int. Array on the GPU.
Contains the vectors ipiv_i of pivot indices returned by getrfBatched.
[in,out]BArray of pointers to type. Each pointer points to an array on the GPU of dimension ldb*nrhs.
On entry, the right hand side matrices B_i. On exit, the solution matrix X_i of each system in the batch.
[in]ldbint. ldb >= n.
The leading dimension of matrices B_i.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of instances (systems) in the batch.

◆ hipblasZgetrsStridedBatched()

hipblasStatus_t hipblasZgetrsStridedBatched ( hipblasHandle_t  handle,
const hipblasOperation_t  trans,
const int  n,
const int  nrhs,
hipblasDoubleComplex A,
const int  lda,
const hipblasStride  strideA,
const int *  ipiv,
const hipblasStride  strideP,
hipblasDoubleComplex B,
const int  ldb,
const hipblasStride  strideB,
int *  info,
const int  batchCount 
)

SOLVER API.

getrsStridedBatched solves a batch of systems of n linear equations on n variables in its factorized forms.

For each instance i in the batch, it solves one of the following systems, depending on the value of trans:

\[ \begin{array}{cl} A_i X_i = B_i & \: \text{not transposed,}\\ A_i^T X_i = B_i & \: \text{transposed, or}\\ A_i^H X_i = B_i & \: \text{conjugate transposed.} \end{array} \]

Matrix \(A_i\) is defined by its triangular factors as returned by getrfStridedBatched.

  • Supported precisions in rocSOLVER : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handlehipblasHandle_t.
[in]transhipblasOperation_t.
Specifies the form of the system of equations of each instance in the batch.
[in]nint. n >= 0.
The order of the system, i.e. the number of columns and rows of all A_i matrices.
[in]nrhsint. nrhs >= 0.
The number of right hand sides, i.e., the number of columns of all the matrices B_i.
[in]Apointer to type. Array on the GPU (the size depends on the value of strideA).
The factors L_i and U_i of the factorization A_i = P_i*L_i*U_i returned by getrfStridedBatched.
[in]ldaint. lda >= n.
The leading dimension of matrices A_i.
[in]strideAhipblasStride.
Stride from the start of one matrix A_i to the next one A_(i+1). There is no restriction for the value of strideA. Normal use case is strideA >= lda*n.
[in]ipivpointer to int. Array on the GPU (the size depends on the value of strideP).
Contains the vectors ipiv_i of pivot indices returned by getrfStridedBatched.
[in]stridePhipblasStride.
Stride from the start of one vector ipiv_i to the next one ipiv_(i+1). There is no restriction for the value of strideP. Normal use case is strideP >= n.
[in,out]Bpointer to type. Array on the GPU (size depends on the value of strideB).
On entry, the right hand side matrices B_i. On exit, the solution matrix X_i of each system in the batch.
[in]ldbint. ldb >= n.
The leading dimension of matrices B_i.
[in]strideBhipblasStride.
Stride from the start of one matrix B_i to the next one B_(i+1). There is no restriction for the value of strideB. Normal use case is strideB >= ldb*nrhs.
[out]infopointer to a int on the host.
If info = 0, successful exit. If info = j < 0, the argument at position -j is invalid.
[in]batchCountint. batchCount >= 0.
Number of instances (systems) in the batch.

◆ hipblasZhbmv()

hipblasStatus_t hipblasZhbmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy 
)

BLAS Level 2 API.

hbmv performs the matrix-vector operations

y := alpha*A*x + beta*y

where alpha and beta are scalars, x and y are n element vectors and A is an n by n Hermitian band matrix, with k super-diagonals.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is being supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is being supplied.
[in]n[int] the order of the matrix A.
[in]k[int] the number of super-diagonals of the matrix A. Must be >= 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing matrix A. Of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The leading (k + 1) by n part of A must contain the upper triangular band part of the Hermitian matrix, with the leading diagonal in row (k + 1), the first super-diagonal on the RHS of row k, etc. The top left k by x triangle of A will not be referenced. Ex (upper, lda = n = 4, k = 1): A Represented matrix (0,0) (5,9) (6,8) (7,7) (1, 0) (5, 9) (0, 0) (0, 0) (1,0) (2,0) (3,0) (4,0) (5,-9) (2, 0) (6, 8) (0, 0) (0,0) (0,0) (0,0) (0,0) (0, 0) (6,-8) (3, 0) (7, 7) (0,0) (0,0) (0,0) (0,0) (0, 0) (0, 0) (7,-7) (4, 0)

if uplo == HIPBLAS_FILL_MODE_LOWER: The leading (k + 1) by n part of A must contain the lower triangular band part of the Hermitian matrix, with the leading diagonal in row (1), the first sub-diagonal on the LHS of row 2, etc. The bottom right k by k triangle of A will not be referenced. Ex (lower, lda = 2, n = 4, k = 1): A Represented matrix (1,0) (2,0) (3,0) (4,0) (1, 0) (5,-9) (0, 0) (0, 0) (5,9) (6,8) (7,7) (0,0) (5, 9) (2, 0) (6,-8) (0, 0) (0, 0) (6, 8) (3, 0) (7,-7) (0, 0) (0, 0) (7, 7) (4, 0)

As a Hermitian matrix, the imaginary part of the main diagonal of A will not be referenced and is assumed to be == 0.

Parameters
[in]lda[int] specifies the leading dimension of A. must be >= k + 1
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasZhbmvBatched()

hipblasStatus_t hipblasZhbmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

hbmvBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian band matrix with k super-diagonals, for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is being supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is being supplied.
[in]n[int] the order of each matrix A_i.
[in]k[int] the number of super-diagonals of each matrix A_i. Must be >= 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The leading (k + 1) by n part of each A_i must contain the upper triangular band part of the Hermitian matrix, with the leading diagonal in row (k + 1), the first super-diagonal on the RHS of row k, etc. The top left k by x triangle of each A_i will not be referenced. Ex (upper, lda = n = 4, k = 1): A Represented matrix (0,0) (5,9) (6,8) (7,7) (1, 0) (5, 9) (0, 0) (0, 0) (1,0) (2,0) (3,0) (4,0) (5,-9) (2, 0) (6, 8) (0, 0) (0,0) (0,0) (0,0) (0,0) (0, 0) (6,-8) (3, 0) (7, 7) (0,0) (0,0) (0,0) (0,0) (0, 0) (0, 0) (7,-7) (4, 0)

if uplo == HIPBLAS_FILL_MODE_LOWER: The leading (k + 1) by n part of each A_i must contain the lower triangular band part of the Hermitian matrix, with the leading diagonal in row (1), the first sub-diagonal on the LHS of row 2, etc. The bottom right k by k triangle of each A_i will not be referenced. Ex (lower, lda = 2, n = 4, k = 1): A Represented matrix (1,0) (2,0) (3,0) (4,0) (1, 0) (5,-9) (0, 0) (0, 0) (5,9) (6,8) (7,7) (0,0) (5, 9) (2, 0) (6,-8) (0, 0) (0, 0) (6, 8) (3, 0) (7,-7) (0, 0) (0, 0) (7, 7) (4, 0)

As a Hermitian matrix, the imaginary part of the main diagonal of each A_i will not be referenced and is assumed to be == 0.

Parameters
[in]lda[int] specifies the leading dimension of each A_i. must be >= max(1, n)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of y.
[in]batchCount[int] number of instances in the batch.

◆ hipblasZhbmvStridedBatched()

hipblasStatus_t hipblasZhbmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

hbmvStridedBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian band matrix with k super-diagonals, for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is being supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is being supplied.
[in]n[int] the order of each matrix A_i.
[in]k[int] the number of super-diagonals of each matrix A_i. Must be >= 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array pointing to the first matrix A_1. Each A_i is of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The leading (k + 1) by n part of each A_i must contain the upper triangular band part of the Hermitian matrix, with the leading diagonal in row (k + 1), the first super-diagonal on the RHS of row k, etc. The top left k by x triangle of each A_i will not be referenced. Ex (upper, lda = n = 4, k = 1): A Represented matrix (0,0) (5,9) (6,8) (7,7) (1, 0) (5, 9) (0, 0) (0, 0) (1,0) (2,0) (3,0) (4,0) (5,-9) (2, 0) (6, 8) (0, 0) (0,0) (0,0) (0,0) (0,0) (0, 0) (6,-8) (3, 0) (7, 7) (0,0) (0,0) (0,0) (0,0) (0, 0) (0, 0) (7,-7) (4, 0)

if uplo == HIPBLAS_FILL_MODE_LOWER: The leading (k + 1) by n part of each A_i must contain the lower triangular band part of the Hermitian matrix, with the leading diagonal in row (1), the first sub-diagonal on the LHS of row 2, etc. The bottom right k by k triangle of each A_i will not be referenced. Ex (lower, lda = 2, n = 4, k = 1): A Represented matrix (1,0) (2,0) (3,0) (4,0) (1, 0) (5,-9) (0, 0) (0, 0) (5,9) (6,8) (7,7) (0,0) (5, 9) (2, 0) (6,-8) (0, 0) (0, 0) (6, 8) (3, 0) (7,-7) (0, 0) (0, 0) (7, 7) (4, 0)

As a Hermitian matrix, the imaginary part of the main diagonal of each A_i will not be referenced and is assumed to be == 0.

Parameters
[in]lda[int] specifies the leading dimension of each A_i. must be >= max(1, n)
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xdevice array pointing to the first vector y_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1)
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array pointing to the first vector y_1.
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZhemm()

hipblasStatus_t hipblasZhemm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex BP,
int  ldb,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc 
)

BLAS Level 3 API.

hemm performs one of the matrix-matrix operations:

C := alpha*A*B + beta*C if side == HIPBLAS_SIDE_LEFT, C := alpha*B*A + beta*C if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B and C are m by n matrices, and A is a Hermitian matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C := alpha*A*B + beta*C HIPBLAS_SIDE_RIGHT: C := alpha*B*A + beta*C
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix
[in]n[int] n specifies the number of rows of B and C. n >= 0.
[in]k[int] n specifies the number of columns of B and C. k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A and B are not referenced.
[in]APpointer storing matrix A on the GPU. A is m by m if side == HIPBLAS_SIDE_LEFT A is n by n if side == HIPBLAS_SIDE_RIGHT Only the upper/lower triangular part is accessed. The imaginary component of the diagonal elements is not used.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPpointer storing matrix B on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m )

◆ hipblasZhemmBatched()

hipblasStatus_t hipblasZhemmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  BP[],
int  ldb,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

hemmBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a Hermitian matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]n[int] n specifies the number of rows of B_i and C_i. n >= 0.
[in]k[int] k specifies the number of columns of B_i and C_i. k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT Only the upper/lower triangular part is accessed. The imaginary component of the diagonal elements is not used.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPdevice array of device pointers storing each matrix B_i on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C_i need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C_i. ldc >= max( 1, m )
[in]batchCount[int] number of instances in the batch.

◆ hipblasZhemmStridedBatched()

hipblasStatus_t hipblasZhemmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex BP,
int  ldb,
hipblasStride  strideB,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

hemmStridedBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a Hermitian matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]n[int] n specifies the number of rows of B_i and C_i. n >= 0.
[in]k[int] k specifies the number of columns of B_i and C_i. k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice pointer to first matrix A_1 A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT Only the upper/lower triangular part is accessed. The imaginary component of the diagonal elements is not used.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPdevice pointer to first matrix B_1 of dimension (ldb, n) on the GPU
[in]ldb[int] ldb specifies the first dimension of B_i. if side = HIPBLAS_OP_N, ldb >= max( 1, m ), otherwise ldb >= max( 1, n ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice pointer to first matrix C_1 of dimension (ldc, n) on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m )
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasZhemv()

hipblasStatus_t hipblasZhemv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy 
)

BLAS Level 2 API.

hemv performs one of the matrix-vector operations

y := alpha*A*x + beta*y

where alpha and beta are scalars, x and y are n element vectors and A is an n by n Hermitian matrix.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of the Hermitian matrix A is supplied. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of the Hermitian matrix A is supplied.
[in]n[int] the order of the matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing matrix A. Of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A must contain the upper triangular part of a Hermitian matrix. The lower triangular part of A will not be referenced. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A must contain the lower triangular part of a Hermitian matrix. The upper triangular part of A will not be referenced. As a Hermitian matrix, the imaginary part of the main diagonal of A will not be referenced and is assumed to be == 0.
[in]lda[int] specifies the leading dimension of A. must be >= max(1, n)
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasZhemvBatched()

hipblasStatus_t hipblasZhemvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

hemvBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian matrix, for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of the Hermitian matrix A is supplied. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of the Hermitian matrix A is supplied.
[in]n[int] the order of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i must contain the upper triangular part of a Hermitian matrix. The lower triangular part of each A_i will not be referenced. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i must contain the lower triangular part of a Hermitian matrix. The upper triangular part of each A_i will not be referenced. As a Hermitian matrix, the imaginary part of the main diagonal of each A_i will not be referenced and is assumed to be == 0.
[in]lda[int] specifies the leading dimension of each A_i. must be >= max(1, n)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of y.
[in]batchCount[int] number of instances in the batch.

◆ hipblasZhemvStridedBatched()

hipblasStatus_t hipblasZhemvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

hemvStridedBatched performs one of the matrix-vector operations

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian matrix, for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of the Hermitian matrix A is supplied. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of the Hermitian matrix A is supplied.
[in]n[int] the order of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice array of device pointers storing each matrix A_i of dimension (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i must contain the upper triangular part of a Hermitian matrix. The lower triangular part of each A_i will not be referenced. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i must contain the lower triangular part of a Hermitian matrix. The upper triangular part of each A_i will not be referenced. As a Hermitian matrix, the imaginary part of the main diagonal of each A_i will not be referenced and is assumed to be == 0.
[in]lda[int] specifies the leading dimension of each A_i. must be >= max(1, n)
[in]strideA[hipblasStride] stride from the start of one (A_i) to the next (A_i+1)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZher()

hipblasStatus_t hipblasZher ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasDoubleComplex AP,
int  lda 
)

BLAS Level 2 API.

her performs the matrix-vector operations

A := A + alpha*x*x**H

where alpha is a real scalar, x is a vector, and A is an n by n Hermitian matrix.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in A. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in A.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing the specified triangular portion of the Hermitian matrix A. Of size (lda * n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the Hermitian matrix A is supplied. The lower triangluar portion will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the Hermitian matrix A is supplied. The upper triangular portion will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of A. Must be at least max(1, n).

◆ hipblasZher2()

hipblasStatus_t hipblasZher2 ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex y,
int  incy,
hipblasDoubleComplex AP,
int  lda 
)

BLAS Level 2 API.

her2 performs the matrix-vector operations

A := A + alpha*x*y**H + conj(alpha)*y*x**H

where alpha is a complex scalar, x and y are vectors, and A is an n by n Hermitian matrix.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing the specified triangular portion of the Hermitian matrix A. Of size (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the Hermitian matrix A is supplied. The lower triangular portion of A will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the Hermitian matrix A is supplied. The upper triangular portion of A will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of A. Must be at least max(lda, 1).

◆ hipblasZher2Batched()

hipblasStatus_t hipblasZher2Batched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex *const  y[],
int  incy,
hipblasDoubleComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

her2Batched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**H + conj(alpha)*y_i*x_i**H

where alpha is a complex scalar, x_i and y_i are vectors, and A_i is an n by n Hermitian matrix for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in,out]APdevice array of device pointers storing the specified triangular portion of each Hermitian matrix A_i of size (lda, n). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The lower triangular portion of each A_i will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The upper triangular portion of each A_i will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of each A_i. Must be at least max(lda, 1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZher2k()

hipblasStatus_t hipblasZher2k ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex BP,
int  ldb,
const double *  beta,
hipblasDoubleComplex CP,
int  ldc 
)

BLAS Level 3 API.

her2k performs one of the matrix-matrix operations for a Hermitian rank-2k update

C := alpha*op( A )*op( B )^H + conj(alpha)*op( B )*op( A )^H + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrices, and C is a n x n Hermitian matrix stored as either upper or lower.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^H, op( B ) = B^H,  and A and B are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op( A ) = A^H, op( B ) = B^H HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasZher2kBatched()

hipblasStatus_t hipblasZher2kBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  BP[],
int  ldb,
const double *  beta,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

her2kBatched performs a batch of the matrix-matrix operations for a Hermitian rank-2k update

C_i := alpha*op( A_i )*op( B_i )^H + conj(alpha)*op( B_i )*op( A_i )^H + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrices, and C_i is a n x n Hermitian matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^H, op( B_i ) = B_i^H,  and A_i and B_i are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op(A) = A^H HIPBLAS_OP_N: op(A) = A
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZher2kStridedBatched()

hipblasStatus_t hipblasZher2kStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex BP,
int  ldb,
hipblasStride  strideB,
const double *  beta,
hipblasDoubleComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

her2kStridedBatched performs a batch of the matrix-matrix operations for a Hermitian rank-2k update

C_i := alpha*op( A_i )*op( B_i )^H + conj(alpha)*op( B_i )*op( A_i )^H + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrices, and C_i is a n x n Hermitian matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^H, op( B_i ) = B_i^H,  and A_i and B_i are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op( A_i ) = A_i^H, op( B_i ) = B_i^H HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZher2StridedBatched()

hipblasStatus_t hipblasZher2StridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

her2StridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**H + conj(alpha)*y_i*x_i**H

where alpha is a complex scalar, x_i and y_i are vectors, and A_i is an n by n Hermitian matrix for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the stride between the beginning of one vector (x_i) and the next (x_i+1).
[in]ydevice pointer pointing to the first vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] specifies the stride between the beginning of one vector (y_i) and the next (y_i+1).
[in,out]APdevice pointer pointing to the first matrix (A_1). Stores the specified triangular portion of each Hermitian matrix A_i. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The lower triangular portion of each A_i will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The upper triangular portion of each A_i will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of each A_i. Must be at least max(lda, 1).
[in]strideA[hipblasStride] specifies the stride between the beginning of one matrix (A_i) and the next (A_i+1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZherBatched()

hipblasStatus_t hipblasZherBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const hipblasDoubleComplex *const  x[],
int  incx,
hipblasDoubleComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

herBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**H

where alpha is a real scalar, x_i is a vector, and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in A. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in A.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing the specified triangular portion of each Hermitian matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The lower triangular portion of each A_i will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The upper triangular portion of each A_i will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of each A_i. Must be at least max(1, n).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZherk()

hipblasStatus_t hipblasZherk ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const hipblasDoubleComplex AP,
int  lda,
const double *  beta,
hipblasDoubleComplex CP,
int  ldc 
)

BLAS Level 3 API.

herk performs one of the matrix-matrix operations for a Hermitian rank-k update

C := alpha*op( A )*op( A )^H + beta*C

where alpha and beta are scalars, op(A) is an n by k matrix, and C is a n x n Hermitian matrix stored as either upper or lower.

op( A ) = A,  and A is n by k if transA == HIPBLAS_OP_N
op( A ) = A^H and A is k by n if transA == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op(A) = A^H HIPBLAS_ON_N: op(A) = A
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if transA = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasZherkBatched()

hipblasStatus_t hipblasZherkBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const double *  beta,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

herkBatched performs a batch of the matrix-matrix operations for a Hermitian rank-k update

C_i := alpha*op( A_i )*op( A_i )^H + beta*C_i

where alpha and beta are scalars, op(A) is an n by k matrix, and C_i is a n x n Hermitian matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^H and A_i is k by n if transA == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op(A) = A^H HIPBLAS_OP_N: op(A) = A
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZherkStridedBatched()

hipblasStatus_t hipblasZherkStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const double *  alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const double *  beta,
hipblasDoubleComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

herkStridedBatched performs a batch of the matrix-matrix operations for a Hermitian rank-k update

C_i := alpha*op( A_i )*op( A_i )^H + beta*C_i

where alpha and beta are scalars, op(A) is an n by k matrix, and C_i is a n x n Hermitian matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^H and A_i is k by n if transA == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op(A) = A^H HIPBLAS_OP_N: op(A) = A
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZherkx()

hipblasStatus_t hipblasZherkx ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex BP,
int  ldb,
const double *  beta,
hipblasDoubleComplex CP,
int  ldc 
)

BLAS Level 3 API.

herkx performs one of the matrix-matrix operations for a Hermitian rank-k update

C := alpha*op( A )*op( B )^H + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrices, and C is a n x n Hermitian matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A )*op( B )^T will be Hermitian.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^H, op( B ) = B^H,  and A and B are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op( A ) = A^H, op( B ) = B^H HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasZherkxBatched()

hipblasStatus_t hipblasZherkxBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  BP[],
int  ldb,
const double *  beta,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

herkxBatched performs a batch of the matrix-matrix operations for a Hermitian rank-k update

C_i := alpha*op( A_i )*op( B_i )^H + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrices, and C_i is a n x n Hermitian matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A )*op( B )^T will be Hermitian.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^H, op( B_i ) = B_i^H,  and A_i and B_i are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op(A) = A^H HIPBLAS_OP_N: op(A) = A
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZherkxStridedBatched()

hipblasStatus_t hipblasZherkxStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex BP,
int  ldb,
hipblasStride  strideB,
const double *  beta,
hipblasDoubleComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

herkxStridedBatched performs a batch of the matrix-matrix operations for a Hermitian rank-k update

C_i := alpha*op( A_i )*op( B_i )^H + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrices, and C_i is a n x n Hermitian matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A )*op( B )^T will be Hermitian.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^H, op( B_i ) = B_i^H,  and A_i and B_i are k by n if trans == HIPBLAS_OP_C
  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_C: op( A_i ) = A_i^H, op( B_i ) = B_i^H HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU. The imaginary component of the diagonal elements are not used but are set to zero unless quick return.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZherStridedBatched()

hipblasStatus_t hipblasZherStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

herStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**H

where alpha is a real scalar, x_i is a vector, and A_i is an n by n Hermitian matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in A. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in A.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in,out]APdevice array of device pointers storing the specified triangular portion of each Hermitian matrix A_i. Points to the first matrix (A_1). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The lower triangular portion of each A_i will not be touched. if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The upper triangular portion of each A_i will not be touched. Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZhpmv()

hipblasStatus_t hipblasZhpmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy 
)

BLAS Level 2 API.

hpmv performs the matrix-vector operation

y := alpha*A*x + beta*y

where alpha and beta are scalars, x and y are n element vectors and A is an n by n Hermitian matrix, supplied in packed form (see description below).

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of the Hermitian matrix A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of the Hermitian matrix A is supplied in AP.
[in]n[int] the order of the matrix A, must be >= 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer storing the packed version of the specified triangular portion of the Hermitian matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,1), (4,0), (3,2), (5,-1), (6,0)] (3,-2) (5, 1) (6, 0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,-1), (3,-2), (4,0), (5,1), (6,0)] (3,-2) (5, 1) (6, 0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasZhpmvBatched()

hipblasStatus_t hipblasZhpmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

hpmvBatched performs the matrix-vector operation

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian matrix, supplied in packed form (see description below), for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of each Hermitian matrix A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of each Hermitian matrix A_i is supplied in AP.
[in]n[int] the order of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer of device pointers storing the packed version of the specified triangular portion of each Hermitian matrix A_i. Each A_i is of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that each AP_i contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,1), (4,0), (3,2), (5,-1), (6,0)] (3,-2) (5, 1) (6, 0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that each AP_i contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,-1), (3,-2), (4,0), (5,1), (6,0)] (3,-2) (5, 1) (6, 0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of y.
[in]batchCount[int] number of instances in the batch.

◆ hipblasZhpmvStridedBatched()

hipblasStatus_t hipblasZhpmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
hipblasStride  strideA,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

hpmvStridedBatched performs the matrix-vector operation

y_i := alpha*A_i*x_i + beta*y_i

where alpha and beta are scalars, x_i and y_i are n element vectors and A_i is an n by n Hermitian matrix, supplied in packed form (see description below), for each batch in i = [1, batchCount].

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: the upper triangular part of each Hermitian matrix A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: the lower triangular part of each Hermitian matrix A_i is supplied in AP.
[in]n[int] the order of each matrix A_i.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]APdevice pointer pointing to the beginning of the first matrix (AP_1). Stores the packed version of the specified triangular portion of each Hermitian matrix AP_i of size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that each AP_i contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,1), (4,0), (3,2), (5,-1), (6,0)] (3,-2) (5, 1) (6, 0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that each AP_i contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (3, 2) (2,-1) (4, 0) (5,-1) --—> [(1,0), (2,-1), (3,-2), (4,0), (5,1), (6,0)] (3,-2) (5, 1) (6, 0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]strideA[hipblasStride] stride from the start of one matrix (AP_i) and the next one (AP_i+1).
[in]xdevice array pointing to the beginning of the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in]betadevice pointer or host pointer to scalar beta.
[in,out]ydevice array pointing to the beginning of the first vector (y_1).
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZhpr()

hipblasStatus_t hipblasZhpr ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasDoubleComplex AP 
)

BLAS Level 2 API.

hpr performs the matrix-vector operations

A := A + alpha*x*x**H

where alpha is a real scalar, x is a vector, and A is an n by n Hermitian matrix, supplied in packed form.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in AP.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing the packed version of the specified triangular portion of the Hermitian matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.

◆ hipblasZhpr2()

hipblasStatus_t hipblasZhpr2 ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex y,
int  incy,
hipblasDoubleComplex AP 
)

BLAS Level 2 API.

hpr2 performs the matrix-vector operations

A := A + alpha*x*y**H + conj(alpha)*y*x**H

where alpha is a complex scalar, x and y are vectors, and A is an n by n Hermitian matrix, supplied in packed form.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in AP.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing the packed version of the specified triangular portion of the Hermitian matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the Hermitian matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.

◆ hipblasZhpr2Batched()

hipblasStatus_t hipblasZhpr2Batched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex *const  y[],
int  incy,
hipblasDoubleComplex *const  AP[],
int  batchCount 
)

BLAS Level 2 API.

hpr2Batched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**H + conj(alpha)*y_i*x_i**H

where alpha is a complex scalar, x_i and y_i are vectors, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each Hermitian matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]batchCount[int] number of instances in the batch.

◆ hipblasZhpr2StridedBatched()

hipblasStatus_t hipblasZhpr2StridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
hipblasDoubleComplex AP,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

hpr2StridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*y_i**H + conj(alpha)*y_i*x_i**H

where alpha is a complex scalar, x_i and y_i are vectors, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in]ydevice pointer pointing to the first vector (y_1).
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1).
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each Hermitian matrix A_i. Points to the first matrix (A_1). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZhprBatched()

hipblasStatus_t hipblasZhprBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const hipblasDoubleComplex *const  x[],
int  incx,
hipblasDoubleComplex *const  AP[],
int  batchCount 
)

BLAS Level 2 API.

hprBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**H

where alpha is a real scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each Hermitian matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]batchCount[int] number of instances in the batch.

◆ hipblasZhprStridedBatched()

hipblasStatus_t hipblasZhprStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const double *  alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
hipblasDoubleComplex AP,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

hprStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**H

where alpha is a real scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each Hermitian matrix A_i. Points to the first matrix (A_1). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,1), (3,0), (4,9), (5,3), (6,0)] (4,-9) (5,-3) (6,0) if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each Hermitian matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 3) (1, 0) (2, 1) (4,9) (2,-1) (3, 0) (5,3) --—> [(1,0), (2,-1), (4,-9), (3,0), (5,-3), (6,0)] (4,-9) (5,-3) (6,0) Note that the imaginary part of the diagonal elements are not accessed and are assumed to be 0.
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZrot()

hipblasStatus_t hipblasZrot ( hipblasHandle_t  handle,
int  n,
hipblasDoubleComplex x,
int  incx,
hipblasDoubleComplex y,
int  incy,
const double *  c,
const hipblasDoubleComplex s 
)

BLAS Level 1 API.

rot applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to vectors x and y. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,c,z,sc,dz
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in the x and y vectors.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment between elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment between elements of y.
[in]cdevice pointer or host pointer storing scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer storing scalar sine component of the rotation matrix.

◆ hipblasZrotBatched()

hipblasStatus_t hipblasZrotBatched ( hipblasHandle_t  handle,
int  n,
hipblasDoubleComplex *const  x[],
int  incx,
hipblasDoubleComplex *const  y[],
int  incy,
const double *  c,
const hipblasDoubleComplex s,
int  batchCount 
)

BLAS Level 1 API.

rotBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice array of deivce pointers storing each vector x_i.
[in]incx[int] specifies the increment between elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment between elements of each y_i.
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasZrotg()

hipblasStatus_t hipblasZrotg ( hipblasHandle_t  handle,
hipblasDoubleComplex a,
hipblasDoubleComplex b,
double *  c,
hipblasDoubleComplex s 
)

BLAS Level 1 API.

rotg creates the Givens rotation matrix for the vector (a b). Scalars c and s and arrays a and b may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice pointer or host pointer to input vector element, overwritten with r.
[in,out]bdevice pointer or host pointer to input vector element, overwritten with z.
[in,out]cdevice pointer or host pointer to cosine element of Givens rotation.
[in,out]sdevice pointer or host pointer sine element of Givens rotation.

◆ hipblasZrotgBatched()

hipblasStatus_t hipblasZrotgBatched ( hipblasHandle_t  handle,
hipblasDoubleComplex *const  a[],
hipblasDoubleComplex *const  b[],
double *const  c[],
hipblasDoubleComplex *const  s[],
int  batchCount 
)

BLAS Level 1 API.

rotgBatched creates the Givens rotation matrix for the batched vectors (a_i b_i), for i = 1, ..., batchCount. a, b, c, and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_DEVICE, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice array of device pointers storing each single input vector element a_i, overwritten with r_i.
[in,out]bdevice array of device pointers storing each single input vector element b_i, overwritten with z_i.
[in,out]cdevice array of device pointers storing each cosine element of Givens rotation for the batch.
[in,out]sdevice array of device pointers storing each sine element of Givens rotation for the batch.
[in]batchCount[int] number of batches (length of arrays a, b, c, and s).

◆ hipblasZrotgStridedBatched()

hipblasStatus_t hipblasZrotgStridedBatched ( hipblasHandle_t  handle,
hipblasDoubleComplex a,
hipblasStride  stridea,
hipblasDoubleComplex b,
hipblasStride  strideb,
double *  c,
hipblasStride  stridec,
hipblasDoubleComplex s,
hipblasStride  strides,
int  batchCount 
)

BLAS Level 1 API.

rotgStridedBatched creates the Givens rotation matrix for the strided batched vectors (a_i b_i), for i = 1, ..., batchCount. a, b, c, and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function blocks the CPU until the GPU has finished and the results are available in host memory. If the pointer mode is set to HIPBLAS_POINTER_MODE_HOST, this function returns immediately and synchronization is required to read the results.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in,out]adevice strided_batched pointer or host strided_batched pointer to first single input vector element a_1, overwritten with r.
[in]stridea[hipblasStride] distance between elements of a in batch (distance between a_i and a_(i + 1))
[in,out]bdevice strided_batched pointer or host strided_batched pointer to first single input vector element b_1, overwritten with z.
[in]strideb[hipblasStride] distance between elements of b in batch (distance between b_i and b_(i + 1))
[in,out]cdevice strided_batched pointer or host strided_batched pointer to first cosine element of Givens rotations c_1.
[in]stridec[hipblasStride] distance between elements of c in batch (distance between c_i and c_(i + 1))
[in,out]sdevice strided_batched pointer or host strided_batched pointer to sine element of Givens rotations s_1.
[in]strides[hipblasStride] distance between elements of s in batch (distance between s_i and s_(i + 1))
[in]batchCount[int] number of batches (length of arrays a, b, c, and s).

◆ hipblasZrotStridedBatched()

hipblasStatus_t hipblasZrotStridedBatched ( hipblasHandle_t  handle,
int  n,
hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
const double *  c,
const hipblasDoubleComplex s,
int  batchCount 
)

BLAS Level 1 API.

rotStridedBatched applies the Givens rotation matrix defined by c=cos(alpha) and s=sin(alpha) to strided batched vectors x_i and y_i, for i = 1, ..., batchCount. Scalars c and s may be stored in either host or device memory, location is specified by calling hipblasSetPointerMode.

  • Supported precisions in rocBLAS : s,d,sc,dz
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] number of elements in each x_i and y_i vectors.
[in,out]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment between elements of each x_i.
[in]stridex[hipblasStride] specifies the increment from the beginning of x_i to the beginning of x_(i+1)
[in,out]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment between elements of each y_i.
[in]stridey[hipblasStride] specifies the increment from the beginning of y_i to the beginning of y_(i+1)
[in]cdevice pointer or host pointer to scalar cosine component of the rotation matrix.
[in]sdevice pointer or host pointer to scalar sine component of the rotation matrix.
[in]batchCount[int] the number of x and y arrays, i.e. the number of batches.

◆ hipblasZscal()

hipblasStatus_t hipblasZscal ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex alpha,
hipblasDoubleComplex x,
int  incx 
)

BLAS Level 1 API.

scal scales each element of vector x with scalar alpha.

x := alpha * x
  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : s,d,c,z,cs,zd
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x.
[in]alphadevice pointer or host pointer for the scalar alpha.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasZscalBatched()

hipblasStatus_t hipblasZscalBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex alpha,
hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 1 API.

scalBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasZscalStridedBatched()

hipblasStatus_t hipblasZscalStridedBatched ( hipblasHandle_t  handle,
int  n,
const hipblasDoubleComplex alpha,
hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 1 API.

scalStridedBatched scales each element of vector x_i with scalar alpha, for i = 1, ... , batchCount.

 x_i := alpha * x_i ,

where (x_i) is the i-th instance of the batch.

  • Supported precisions in rocBLAS : s,d,c,z,cs,zd
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i.
[in]alphahost pointer or device pointer for the scalar alpha.
[in,out]xdevice pointer to the first vector (x_1) in the batch.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in]batchCount[int] specifies the number of batches in x.

◆ hipblasZspr()

hipblasStatus_t hipblasZspr ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasDoubleComplex AP 
)

BLAS Level 2 API.

spr performs the matrix-vector operations

A := A + alpha*x*x**T

where alpha is a scalar, x is a vector, and A is an n by n symmetric matrix, supplied in packed form.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of A is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of A is supplied in AP.
[in]n[int] the number of rows and columns of matrix A, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing the packed version of the specified triangular portion of the symmetric matrix A. Of at least size ((n * (n + 1)) / 2). if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of the symmetric matrix A is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0

◆ hipblasZsprBatched()

hipblasStatus_t hipblasZsprBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  x[],
int  incx,
hipblasDoubleComplex *const  AP[],
int  batchCount 
)

BLAS Level 2 API.

sprBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**T

where alpha is a scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing the packed version of the specified triangular portion of each symmetric matrix A_i of at least size ((n * (n + 1)) / 2). Array is of at least size batchCount. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]batchCount[int] number of instances in the batch.

◆ hipblasZsprStridedBatched()

hipblasStatus_t hipblasZsprStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
hipblasDoubleComplex AP,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

sprStridedBatched performs the matrix-vector operations

A_i := A_i + alpha*x_i*x_i**T

where alpha is a scalar, x_i is a vector, and A_i is an n by n symmetric matrix, supplied in packed form, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' HIPBLAS_FILL_MODE_UPPER: The upper triangular part of each A_i is supplied in AP. HIPBLAS_FILL_MODE_LOWER: The lower triangular part of each A_i is supplied in AP.
[in]n[int] the number of rows and columns of each matrix A_i, must be at least 0.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer pointing to the first vector (x_1).
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1).
[in,out]APdevice pointer storing the packed version of the specified triangular portion of each symmetric matrix A_i. Points to the first A_1. if uplo == HIPBLAS_FILL_MODE_UPPER: The upper triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(0,1) AP(2) = A(1,1), etc. Ex: (HIPBLAS_FILL_MODE_UPPER; n = 4) 1 2 4 7 2 3 5 8 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 4 5 6 9 7 8 9 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The lower triangular portion of each symmetric matrix A_i is supplied. The matrix is compacted so that AP contains the triangular portion column-by-column so that: AP(0) = A(0,0) AP(1) = A(1,0) AP(2) = A(2,1), etc. Ex: (HIPBLAS_FILL_MODE_LOWER; n = 4) 1 2 3 4 2 5 6 7 --—> [1, 2, 3, 4, 5, 6, 7, 8, 9, 0] 3 6 8 9 4 7 9 0
[in]strideA[hipblasStride] stride from the start of one (A_i) and the next (A_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZswap()

hipblasStatus_t hipblasZswap ( hipblasHandle_t  handle,
int  n,
hipblasDoubleComplex x,
int  incx,
hipblasDoubleComplex y,
int  incy 
)

BLAS Level 1 API.

swap interchanges vectors x and y.

y := x; x := y
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in x and y.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.

◆ hipblasZswapBatched()

hipblasStatus_t hipblasZswapBatched ( hipblasHandle_t  handle,
int  n,
hipblasDoubleComplex *const  x[],
int  incx,
hipblasDoubleComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 1 API.

swapBatched interchanges vectors x_i and y_i, for i = 1 , ... , batchCount

y_i := x_i; x_i := y_i
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in,out]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasZswapStridedBatched()

hipblasStatus_t hipblasZswapStridedBatched ( hipblasHandle_t  handle,
int  n,
hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 1 API.

swapStridedBatched interchanges vectors x_i and y_i, for i = 1 , ... , batchCount

y_i := x_i; x_i := y_i
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]n[int] the number of elements in each x_i and y_i.
[in,out]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of x.
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_x is of appropriate size, for a typical case this means stride_x >= n * incx.
[in,out]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment for the elements of y.
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stride_x, however the user should take care to ensure that stride_y is of appropriate size, for a typical case this means stride_y >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch.

◆ hipblasZsymm()

hipblasStatus_t hipblasZsymm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex BP,
int  ldb,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc 
)

BLAS Level 3 API.

symm performs one of the matrix-matrix operations:

C := alpha*A*B + beta*C if side == HIPBLAS_SIDE_LEFT, C := alpha*B*A + beta*C if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B and C are m by n matrices, and A is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C := alpha*A*B + beta*C HIPBLAS_SIDE_RIGHT: C := alpha*B*A + beta*C
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix
[in]m[int] m specifies the number of rows of B and C. m >= 0.
[in]n[int] n specifies the number of columns of B and C. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A and B are not referenced.
[in]APpointer storing matrix A on the GPU. A is m by m if side == HIPBLAS_SIDE_LEFT A is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPpointer storing matrix B on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m )

◆ hipblasZsymmBatched()

hipblasStatus_t hipblasZsymmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  BP[],
int  ldb,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

symmBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]m[int] m specifies the number of rows of B_i and C_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i and C_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]BPdevice array of device pointers storing each matrix B_i on the GPU. Matrix dimension is m by n
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m )
[in]betabeta specifies the scalar beta. When beta is zero then C_i need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU. Matrix dimension is m by n
[in]ldc[int] ldc specifies the first dimension of C_i. ldc >= max( 1, m )
[in]batchCount[int] number of instances in the batch.

◆ hipblasZsymmStridedBatched()

hipblasStatus_t hipblasZsymmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex BP,
int  ldb,
hipblasStride  strideB,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

symmStridedBatched performs a batch of the matrix-matrix operations:

C_i := alpha*A_i*B_i + beta*C_i if side == HIPBLAS_SIDE_LEFT, C_i := alpha*B_i*A_i + beta*C_i if side == HIPBLAS_SIDE_RIGHT,

where alpha and beta are scalars, B_i and C_i are m by n matrices, and A_i is a symmetric matrix stored as either upper or lower.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: C_i := alpha*A_i*B_i + beta*C_i HIPBLAS_SIDE_RIGHT: C_i := alpha*B_i*A_i + beta*C_i
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix
[in]m[int] m specifies the number of rows of B_i and C_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i and C_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i and B_i are not referenced.
[in]APdevice pointer to first matrix A_1 A_i is m by m if side == HIPBLAS_SIDE_LEFT A_i is n by n if side == HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), otherwise lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPdevice pointer to first matrix B_1 of dimension (ldb, n) on the GPU.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m )
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice pointer to first matrix C_1 of dimension (ldc, n) on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, m ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZsymv()

hipblasStatus_t hipblasZsymv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy 
)

BLAS Level 2 API.

symv performs the matrix-vector operation:

y := alpha*A*x + beta*y,

where alpha and beta are scalars, x and y are n element vectors and A should contain an upper or lower triangular n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int]
[in]alphaspecifies the scalar alpha
[in]APpointer storing matrix A on the GPU
[in]lda[int] specifies the leading dimension of A
[in]xpointer storing vector x on the GPU
[in]incx[int] specifies the increment for the elements of x
[in]betaspecifies the scalar beta
[out]ypointer storing vector y on the GPU
[in]incy[int] specifies the increment for the elements of y

◆ hipblasZsymvBatched()

hipblasStatus_t hipblasZsymvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  y[],
int  incy,
int  batchCount 
)

BLAS Level 2 API.

symvBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A a should contain an upper or lower triangular symmetric matrix and the opposing triangular part of A is not referenced

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APdevice array of device pointers storing each matrix A_i
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]xdevice array of device pointers storing each vector x_i
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]betadevice pointer or host pointer to scalar beta
[out]ydevice array of device pointers storing each vector y_i
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]batchCount[int] number of instances in the batch

◆ hipblasZsymvStridedBatched()

hipblasStatus_t hipblasZsymvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex beta,
hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
int  batchCount 
)

BLAS Level 2 API.

symvStridedBatched performs the matrix-vector operation:

y_i := alpha*A_i*x_i + beta*y_i,

where (A_i, x_i, y_i) is the i-th instance of the batch. alpha and beta are scalars, x_i and y_i are vectors and A_i is an n by n symmetric matrix, for i = 1, ..., batchCount. A a should contain an upper or lower triangular symmetric matrix and the opposing triangular part of A is not referenced

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] number of rows and columns of each matrix A_i
[in]alphadevice pointer or host pointer to scalar alpha
[in]APDevice pointer to the first matrix A_1 on the GPU
[in]lda[int] specifies the leading dimension of each matrix A_i
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]xDevice pointer to the first vector x_1 on the GPU
[in]incx[int] specifies the increment for the elements of each vector x_i
[in]stridex[hipblasStride] stride from the start of one vector (x_i) and the next one (x_i+1). There are no restrictions placed on stridex, however the user should take care to ensure that stridex is of appropriate size. This typically means stridex >= n * incx. stridex should be non zero.
[in]betadevice pointer or host pointer to scalar beta
[out]yDevice pointer to the first vector y_1 on the GPU
[in]incy[int] specifies the increment for the elements of each vector y_i
[in]stridey[hipblasStride] stride from the start of one vector (y_i) and the next one (y_i+1). There are no restrictions placed on stridey, however the user should take care to ensure that stridey is of appropriate size. This typically means stridey >= n * incy. stridey should be non zero.
[in]batchCount[int] number of instances in the batch

◆ hipblasZsyr()

hipblasStatus_t hipblasZsyr ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasDoubleComplex AP,
int  lda 
)

BLAS Level 2 API.

syr performs the matrix-vector operations

A := A + alpha*x*x**T

where alpha is a scalar, x is a vector, and A is an n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasZsyr2()

hipblasStatus_t hipblasZsyr2 ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
const hipblasDoubleComplex y,
int  incy,
hipblasDoubleComplex AP,
int  lda 
)

BLAS Level 2 API.

syr2 performs the matrix-vector operations

A := A + alpha*x*y**T + alpha*y*x**T

where alpha is a scalar, x and y are vectors, and A is an n by n symmetric matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.
[in]ydevice pointer storing vector y.
[in]incy[int] specifies the increment for the elements of y.
[in,out]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.

◆ hipblasZsyr2Batched()

hipblasStatus_t hipblasZsyr2Batched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  x[],
int  incx,
const hipblasDoubleComplex *const  y[],
int  incy,
hipblasDoubleComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

syr2Batched performs a batch of matrix-vector operations

A[i] := A[i] + alpha*x[i]*y[i]**T + alpha*y[i]*x[i]**T

where alpha is a scalar, x[i] and y[i] are vectors, and A[i] is a n by n symmetric matrix, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]ydevice array of device pointers storing each vector y_i.
[in]incy[int] specifies the increment for the elements of each y_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasZsyr2k()

hipblasStatus_t hipblasZsyr2k ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex BP,
int  ldb,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc 
)

BLAS Level 3 API.

syr2k performs one of the matrix-matrix operations for a symmetric rank-2k update

C := alpha*(op( A )*op( B )^T + op( B )*op( A )^T) + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^T, op( B ) = B^T,  and A and B are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A ) = A^T, op( B ) = B^T HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A) and op(B). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasZsyr2kBatched()

hipblasStatus_t hipblasZsyr2kBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  BP[],
int  ldb,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syr2kBatched performs a batch of the matrix-matrix operations for a symmetric rank-2k update

C_i := alpha*(op( A_i )*op( B_i )^T + op( B_i )*op( A_i )^T) + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZsyr2kStridedBatched()

hipblasStatus_t hipblasZsyr2kStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex BP,
int  ldb,
hipblasStride  strideB,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syr2kStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-2k update

C_i := alpha*(op( A_i )*op( B_i )^T + op( B_i )*op( A_i )^T) + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZsyr2StridedBatched()

hipblasStatus_t hipblasZsyr2StridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
const hipblasDoubleComplex y,
int  incy,
hipblasStride  stridey,
hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

syr2StridedBatched the matrix-vector operations

A[i] := A[i] + alpha*x[i]*y[i]**T + alpha*y[i]*x[i]**T

where alpha is a scalar, x[i] and y[i] are vectors, and A[i] is a n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of each matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the pointer increment between vectors (x_i) and (x_i+1).
[in]ydevice pointer to the first vector y_1.
[in]incy[int] specifies the increment for the elements of each y_i.
[in]stridey[hipblasStride] specifies the pointer increment between vectors (y_i) and (y_i+1).
[in,out]APdevice pointer to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasZsyrBatched()

hipblasStatus_t hipblasZsyrBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  x[],
int  incx,
hipblasDoubleComplex *const  AP[],
int  lda,
int  batchCount 
)

BLAS Level 2 API.

syrBatched performs a batch of matrix-vector operations

A[i] := A[i] + alpha*x[i]*x[i]**T

where alpha is a scalar, x is an array of vectors, and A is an array of n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in,out]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasZsyrk()

hipblasStatus_t hipblasZsyrk ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc 
)

BLAS Level 3 API.

syrk performs one of the matrix-matrix operations for a symmetric rank-k update

C := alpha*op( A )*op( A )^T + beta*C

where alpha and beta are scalars, op(A) is an n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower.

op( A ) = A, and A is n by k if transA == HIPBLAS_OP_N
op( A ) = A^T and A is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if transA = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasZsyrkBatched()

hipblasStatus_t hipblasZsyrkBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syrkBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( A_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) is an n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^T and A_i is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZsyrkStridedBatched()

hipblasStatus_t hipblasZsyrkStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syrkStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( A_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) is an n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower.

op( A_i ) = A_i, and A_i is n by k if transA == HIPBLAS_OP_N
op( A_i ) = A_i^T and A_i is k by n if transA == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op(A) = A^T HIPBLAS_OP_N: op(A) = A HIPBLAS_OP_C: op(A) = A^T

HIPBLAS_OP_C is not supported for complex types, see cherk and zherk.

Parameters
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when transA is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if transA = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU. on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZsyrkx()

hipblasStatus_t hipblasZsyrkx ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex BP,
int  ldb,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc 
)

BLAS Level 3 API.

syrkx performs one of the matrix-matrix operations for a symmetric rank-k update

C := alpha*op( A )*op( B )^T + beta*C

where alpha and beta are scalars, op(A) and op(B) are n by k matrix, and C is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A )*op( B )^T will be symmetric.

op( A ) = A, op( B ) = B, and A and B are n by k if trans == HIPBLAS_OP_N
op( A ) = A^T, op( B ) = B^T,  and A and B are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A ) = A^T, op( B ) = B^T HIPBLAS_OP_N: op( A ) = A, op( B ) = B
[in]n[int] n specifies the number of rows and columns of C. n >= 0.
[in]k[int] k specifies the number of columns of op(A) and op(B). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APpointer storing matrix A on the GPU. Martrix dimension is ( lda, k ) when if trans = HIPBLAS_OP_N, otherwise (lda, n) only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPpointer storing matrix B on the GPU. Martrix dimension is ( ldb, k ) when if trans = HIPBLAS_OP_N, otherwise (ldb, n) only the upper/lower triangular part is accessed.
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPpointer storing matrix C on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).

◆ hipblasZsyrkxBatched()

hipblasStatus_t hipblasZsyrkxBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  BP[],
int  ldb,
const hipblasDoubleComplex beta,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

BLAS Level 3 API.

syrkxBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( B_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A_i )*op( B_i )^T will be symmetric.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APdevice array of device pointers storing each matrix_i A of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]BPdevice array of device pointers storing each matrix_i B of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPdevice array of device pointers storing each matrix C_i on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZsyrkxStridedBatched()

hipblasStatus_t hipblasZsyrkxStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
int  n,
int  k,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex BP,
int  ldb,
hipblasStride  strideB,
const hipblasDoubleComplex beta,
hipblasDoubleComplex CP,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

BLAS Level 3 API.

syrkxStridedBatched performs a batch of the matrix-matrix operations for a symmetric rank-k update

C_i := alpha*op( A_i )*op( B_i )^T + beta*C_i

where alpha and beta are scalars, op(A_i) and op(B_i) are n by k matrix, and C_i is a symmetric n x n matrix stored as either upper or lower. This routine should only be used when the caller can guarantee that the result of op( A_i )*op( B_i )^T will be symmetric.

op( A_i ) = A_i, op( B_i ) = B_i, and A_i and B_i are n by k if trans == HIPBLAS_OP_N
op( A_i ) = A_i^T, op( B_i ) = B_i^T,  and A_i and B_i are k by n if trans == HIPBLAS_OP_T
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: C_i is an upper triangular matrix HIPBLAS_FILL_MODE_LOWER: C_i is a lower triangular matrix
[in]transA[hipblasOperation_t] HIPBLAS_OP_T: op( A_i ) = A_i^T, op( B_i ) = B_i^T HIPBLAS_OP_N: op( A_i ) = A_i, op( B_i ) = B_i
[in]n[int] n specifies the number of rows and columns of C_i. n >= 0.
[in]k[int] k specifies the number of columns of op(A). k >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and A need not be set before entry.
[in]APDevice pointer to the first matrix A_1 on the GPU of dimension (lda, k) when trans is HIPBLAS_OP_N, otherwise of dimension (lda, n)
[in]lda[int] lda specifies the first dimension of A_i. if trans = HIPBLAS_OP_N, lda >= max( 1, n ), otherwise lda >= max( 1, k ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]BPDevice pointer to the first matrix B_1 on the GPU of dimension (ldb, k) when trans is HIPBLAS_OP_N, otherwise of dimension (ldb, n)
[in]ldb[int] ldb specifies the first dimension of B_i. if trans = HIPBLAS_OP_N, ldb >= max( 1, n ), otherwise ldb >= max( 1, k ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]betabeta specifies the scalar beta. When beta is zero then C need not be set before entry.
[in]CPDevice pointer to the first matrix C_1 on the GPU.
[in]ldc[int] ldc specifies the first dimension of C. ldc >= max( 1, n ).
[in,out]strideC[hipblasStride] stride from the start of one matrix (C_i) and the next one (C_i+1)
[in]batchCount[int] number of instances in the batch.

◆ hipblasZsyrStridedBatched()

hipblasStatus_t hipblasZsyrStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
int  batchCount 
)

BLAS Level 2 API.

syrStridedBatched performs the matrix-vector operations

A[i] := A[i] + alpha*x[i]*x[i]**T

where alpha is a scalar, vectors, and A is an array of n by n symmetric matrices, for i = 1 , ... , batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]n[int] the number of rows and columns of each matrix A.
[in]alphadevice pointer or host pointer to scalar alpha.
[in]xdevice pointer to the first vector x_1.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the pointer increment between vectors (x_i) and (x_i+1).
[in,out]APdevice pointer to the first matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i.
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in]batchCount[int] number of instances in the batch

◆ hipblasZtbmv()

hipblasStatus_t hipblasZtbmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const hipblasDoubleComplex AP,
int  lda,
hipblasDoubleComplex x,
int  incx 
)

BLAS Level 2 API.

tbmv performs one of the matrix-vector operations

x := A*x      or
x := A**T*x   or
x := A**H*x,

x is a vectors and A is a banded m by m matrix (see description below).

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether matrix A is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of A is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of A's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by A.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of the matrix A. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of the matrix A. k must satisfy k > 0 && k < lda.
[in]APdevice pointer storing banded triangular matrix A. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of A. lda must satisfy lda > k.
[in,out]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasZtbmvBatched()

hipblasStatus_t hipblasZtbmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const hipblasDoubleComplex *const  AP[],
int  lda,
hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tbmvBatched performs one of the matrix-vector operations

x_i := A_i*x_i      or
x_i := A_i**T*x_i   or
x_i := A_i**H*x_i,

where (A_i, x_i) is the i-th instance of the batch. x_i is a vector and A_i is an m by m matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether each matrix A_i is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of each A_i is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of each A_i's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by each A_i.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of each matrix A_i. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of each matrix A_i. k must satisfy k > 0 && k < lda.
[in]APdevice array of device pointers storing each banded triangular matrix A_i. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of each A_i. lda must satisfy lda > k.
[in,out]xdevice array of device pointer storing each vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasZtbmvStridedBatched()

hipblasStatus_t hipblasZtbmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  k,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tbmvStridedBatched performs one of the matrix-vector operations

x_i := A_i*x_i      or
x_i := A_i**T*x_i   or
x_i := A_i**H*x_i,

where (A_i, x_i) is the i-th instance of the batch. x_i is a vector and A_i is an m by m matrix, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper banded triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower banded triangular matrix.
[in]transA[hipblasOperation_t] indicates whether each matrix A_i is tranposed (conjugated) or not.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: The main diagonal of each A_i is assumed to consist of only 1's and is not referenced. HIPBLAS_DIAG_NON_UNIT: No assumptions are made of each A_i's main diagonal.
[in]m[int] the number of rows and columns of the matrix represented by each A_i.
[in]k[int] if uplo == HIPBLAS_FILL_MODE_UPPER, k specifies the number of super-diagonals of each matrix A_i. if uplo == HIPBLAS_FILL_MODE_LOWER, k specifies the number of sub-diagonals of each matrix A_i. k must satisfy k > 0 && k < lda.
[in]APdevice array to the first matrix A_i of the batch. Stores each banded triangular matrix A_i. if uplo == HIPBLAS_FILL_MODE_UPPER: The matrix represented is an upper banded triangular matrix with the main diagonal and k super-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the k'th row, the first super diagonal resides on the RHS of the k-1'th row, etc, with the k'th diagonal on the RHS of the 0'th row. Ex: (HIPBLAS_FILL_MODE_UPPER; m = 5; k = 2) 1 6 9 0 0 0 0 9 8 7 0 2 7 8 0 0 6 7 8 9 0 0 3 8 7 -—> 1 2 3 4 5 0 0 0 4 9 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 if uplo == HIPBLAS_FILL_MODE_LOWER: The matrix represnted is a lower banded triangular matrix with the main diagonal and k sub-diagonals, everything else can be assumed to be 0. The matrix is compacted so that the main diagonal resides on the 0'th row, working up to the k'th diagonal residing on the LHS of the k'th row. Ex: (HIPBLAS_FILL_MODE_LOWER; m = 5; k = 2) 1 0 0 0 0 1 2 3 4 5 6 2 0 0 0 6 7 8 9 0 9 7 3 0 0 -—> 9 8 7 0 0 0 8 8 4 0 0 0 0 0 0 0 0 7 9 5 0 0 0 0 0
[in]lda[int] specifies the leading dimension of each A_i. lda must satisfy lda > k.
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in,out]xdevice array to the first vector x_i of the batch.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the start of one x_i matrix to the next x_(i + 1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZtbsv()

hipblasStatus_t hipblasZtbsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const hipblasDoubleComplex AP,
int  lda,
hipblasDoubleComplex x,
int  incx 
)

BLAS Level 2 API.

tbsv solves

 A*x = b or A**T*x = b or A**H*x = b,

where x and b are vectors and A is a banded triangular matrix.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular (i.e. the diagonal elements of A are not used in computations). HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of b. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of A. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of A. k >= 0.
[in]APdevice pointer storing the matrix A in banded format.
[in]lda[int] specifies the leading dimension of A. lda >= (k + 1).
[in,out]xdevice pointer storing input vector b. Overwritten by the output vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasZtbsvBatched()

hipblasStatus_t hipblasZtbsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const hipblasDoubleComplex *const  AP[],
int  lda,
hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tbsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a banded triangular matrix, for i = [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A_i*x_i = b_i HIPBLAS_OP_T: Solves A_i**T*x_i = b_i HIPBLAS_OP_C: Solves A_i**H*x_i = b_i
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of each b_i. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of each A_i. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of each A_i. k >= 0.
[in]APdevice vector of device pointers storing each matrix A_i in banded format.
[in]lda[int] specifies the leading dimension of each A_i. lda >= (k + 1).
[in,out]xdevice vector of device pointers storing each input vector b_i. Overwritten by each output vector x_i.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch.

◆ hipblasZtbsvStridedBatched()

hipblasStatus_t hipblasZtbsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  n,
int  k,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tbsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a banded triangular matrix, for i = [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A_i*x_i = b_i HIPBLAS_OP_T: Solves A_i**T*x_i = b_i HIPBLAS_OP_C: Solves A_i**H*x_i = b_i
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]n[int] n specifies the number of rows of each b_i. n >= 0.
[in]k[int] if(uplo == HIPBLAS_FILL_MODE_UPPER) k specifies the number of super-diagonals of each A_i. if(uplo == HIPBLAS_FILL_MODE_LOWER) k specifies the number of sub-diagonals of each A_i. k >= 0.
[in]APdevice pointer pointing to the first banded matrix A_1.
[in]lda[int] specifies the leading dimension of each A_i. lda >= (k + 1).
[in]strideA[hipblasStride] specifies the distance between the start of one matrix (A_i) and the next (A_i+1).
[in,out]xdevice pointer pointing to the first input vector b_1. Overwritten by output vectors x.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] specifies the distance between the start of one vector (x_i) and the next (x_i+1).
[in]batchCount[int] number of instances in the batch.

◆ hipblasZtpmv()

hipblasStatus_t hipblasZtpmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex AP,
hipblasDoubleComplex x,
int  incx 
)

BLAS Level 2 API.

tpmv performs one of the matrix-vector operations

 x = A*x or x = A**T*x,

where x is an n element vector and A is an n by n unit, or non-unit, upper or lower triangular matrix, supplied in the pack form.

The vector x is overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of A. m >= 0.
[in]APdevice pointer storing matrix A, of dimension at leat ( m * ( m + 1 ) / 2 ). Before entry with uplo = HIPBLAS_FILL_MODE_UPPER, the array A must contain the upper triangular matrix packed sequentially, column by column, so that A[0] contains a_{0,0}, A[1] and A[2] contain a_{0,1} and a_{1, 1} respectively, and so on. Before entry with uplo = HIPBLAS_FILL_MODE_LOWER, the array A must contain the lower triangular matrix packed sequentially, column by column, so that A[0] contains a_{0,0}, A[1] and A[2] contain a_{1,0} and a_{2,0} respectively, and so on. Note that when DIAG = HIPBLAS_DIAG_UNIT, the diagonal elements of A are not referenced, but are assumed to be unity.
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x. incx must not be zero.

◆ hipblasZtpmvBatched()

hipblasStatus_t hipblasZtpmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex *const  AP[],
hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tpmvBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix)

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer storing pointer of matrices A_i, of dimension ( lda, m )
[in]xdevice pointer storing vectors x_i.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasZtpmvStridedBatched()

hipblasStatus_t hipblasZtpmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex AP,
hipblasStride  strideA,
hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tpmvStridedBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix) with strides specifying how to retrieve $x_i$ (resp. $A_i$) from $x_{i-1}$ (resp. $A_i$).

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer of the matrix A_0, of dimension ( lda, m )
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_{i + 1}
[in]xdevice pointer storing the vector x_0.
[in]incx[int] specifies the increment for the elements of one vector x.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_{i + 1}
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasZtpsv()

hipblasStatus_t hipblasZtpsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex AP,
hipblasDoubleComplex x,
int  incx 
)

BLAS Level 2 API.

tpsv solves

 A*x = b or A**T*x = b, or A**H*x = b,

where x and b are vectors and A is a triangular matrix stored in the packed format.

The input vector b is overwritten by the output vector x.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular (i.e. the diagonal elements of A are not used in computations). HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice pointer storing the packed version of matrix A, of dimension >= (n * (n + 1) / 2)
[in,out]xdevice pointer storing vector b on input, overwritten by x on output.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasZtpsvBatched()

hipblasStatus_t hipblasZtpsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex *const  AP[],
hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

tpsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i, or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a triangular matrix stored in the packed format, for i in [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice array of device pointers storing the packed versions of each matrix A_i, of dimension >= (n * (n + 1) / 2)
[in,out]xdevice array of device pointers storing each input vector b_i, overwritten by x_i on output.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasZtpsvStridedBatched()

hipblasStatus_t hipblasZtpsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex AP,
hipblasStride  strideA,
hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

tpsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i, or A_i**H*x_i = b_i,

where x_i and b_i are vectors and A_i is a triangular matrix stored in the packed format, for i in [1, batchCount].

The input vectors b_i are overwritten by the output vectors x_i.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: Solves A*x = b HIPBLAS_OP_T: Solves A**T*x = b HIPBLAS_OP_C: Solves A**H*x = b
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular (i.e. the diagonal elements of each A_i are not used in computations). HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice pointer pointing to the first packed matrix A_1, of dimension >= (n * (n + 1) / 2)
[in]strideA[hipblasStride] stride from the beginning of one packed matrix (AP_i) and the next (AP_i+1).
[in,out]xdevice pointer pointing to the first input vector b_1. Overwritten by each x_i on output.
[in]incx[int] specifies the increment for the elements of each x_i.
[in]stridex[hipblasStride] stride from the beginning of one vector (x_i) and the next (x_i+1).
[in]batchCount[int] specifies the number of instances in the batch.

◆ hipblasZtrmm()

hipblasStatus_t hipblasZtrmm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasDoubleComplex BP,
int  ldb 
)

BLAS Level 3 API.

#ifndef HIPBLAS_V1 // deprecated

    hipblasStatus_t hipblasStrmm(hipblasHandle_t    handle,
                                 hipblasSideMode_t  side,
                                 hipblasFillMode_t  uplo,
                                 hipblasOperation_t transA,
                                 hipblasDiagType_t  diag,
                                 int                m,
                                 int                n,
                                 const float*       alpha,
                                 const float*       AP,
                                 int                lda,
                                 float*             BP,
                                 int                ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasStrmmOutofplace(hipblasHandle_t    handle,
                                          hipblasSideMode_t  side,
                                          hipblasFillMode_t  uplo,
                                          hipblasOperation_t transA,
                                          hipblasDiagType_t  diag,
                                          int                m,
                                          int                n,
                                          const float*       alpha,
                                          const float*       AP,
                                          int                lda,
                                          const float*       BP,
                                          int                ldb,
                                          float*             CP,
                                          int                ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasDtrmm(hipblasHandle_t    handle,
                                hipblasSideMode_t  side,
                                hipblasFillMode_t  uplo,
                                hipblasOperation_t transA,
                                hipblasDiagType_t  diag,
                                int                m,
                                int                n,
                                const double*      alpha,
                                const double*      AP,
                                int                lda,
                                double*            BP,
                                int                ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasDtrmmOutofplace(hipblasHandle_t    handle,
                                          hipblasSideMode_t  side,
                                          hipblasFillMode_t  uplo,
                                          hipblasOperation_t transA,
                                          hipblasDiagType_t  diag,
                                          int                m,
                                          int                n,
                                          const double*      alpha,
                                          const double*      AP,
                                          int                lda,
                                          const double*      BP,
                                          int                ldb,
                                          double*            CP,
                                          int                ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasCtrmm(hipblasHandle_t       handle,
                                hipblasSideMode_t     side,
                                hipblasFillMode_t     uplo,
                                hipblasOperation_t    transA,
                                hipblasDiagType_t     diag,
                                int                   m,
                                int                   n,
                                const hipblasComplex* alpha,
                                const hipblasComplex* AP,
                                int                   lda,
                                hipblasComplex*       BP,
                                int                   ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasCtrmmOutofplace(hipblasHandle_t       handle,
                                          hipblasSideMode_t     side,
                                          hipblasFillMode_t     uplo,
                                          hipblasOperation_t    transA,
                                          hipblasDiagType_t     diag,
                                          int                   m,
                                          int                   n,
                                          const hipblasComplex* alpha,
                                          const hipblasComplex* AP,
                                          int                   lda,
                                          const hipblasComplex* BP,
                                          int                   ldb,
                                          hipblasComplex*       CP,
                                          int                   ldc);

#endif

#ifndef HIPBLAS_V1 // deprecated

   hipblasStatus_t hipblasZtrmm(hipblasHandle_t             handle,
                                hipblasSideMode_t           side,
                                hipblasFillMode_t           uplo,
                                hipblasOperation_t          transA,
                                hipblasDiagType_t           diag,
                                int                         m,
                                int                         n,
                                const hipblasDoubleComplex* alpha,
                                const hipblasDoubleComplex* AP,
                                int                         lda,
                                hipblasDoubleComplex*       BP,
                                int                         ldb);

#else // available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

   hipblasStatus_t hipblasZtrmmOutofplace(hipblasHandle_t             handle,
                                          hipblasSideMode_t           side,
                                          hipblasFillMode_t           uplo,
                                          hipblasOperation_t          transA,
                                          hipblasDiagType_t           diag,
                                          int                         m,
                                          int                         n,
                                          const hipblasDoubleComplex* alpha,
                                          const hipblasDoubleComplex* AP,
                                          int                         lda,
                                          const hipblasDoubleComplex* BP,
                                          int                         ldb,
                                          hipblasDoubleComplex*       CP,
                                          int                         ldc);

#endif

The deprecated Legacy BLAS in-place trmm performs one of the matrix-matrix operations:

B := alpha*op( A )*B,   or
B := alpha*B*op( A ),

The new trmm performs one of the matrix-matrix operations:

C := alpha*op( A )*B,   or
C := alpha*B*op( A ),

The in-place functionality is still available in the new trmmm by setting pointer C equal to pointer B, and ldc equal to ldb.

alpha  is a scalar,  B  is an m by n matrix, C  is an m by n matrix,  A  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A )  is one  of

op( A ) = A     or
op( A ) = A^T   or
op( A ) = A^H.

When uplo == rocblas_fill_upper the  leading  k by k
upper triangular part of the array  A must contain the upper
triangular matrix and the strictly lower triangular part of
A is not referenced. Here k is m when side == rocblas_side_left
and is n when side == rocblas_side_right.

When uplo == rocblas_fill_lower the  leading  k by k
lower triangular part of the array  A must contain the lower
triangular matrix  and the strictly upper triangular part of
A is not referenced. Here k is m when  side == rocblas_side_left
and is n when side == rocblas_side_right.

Note that when  diag == rocblas_diagonal_unit  the diagonal elements of
A  are not referenced either,  but are assumed to be  unity.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A) multiplies B from the left or right as follows: HIPBLAS_SIDE_LEFT: B := alpha*op( A )*B. HIPBLAS_SIDE_RIGHT: B := alpha*B*op( A ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A is unit triangular as follows: HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B. m >= 0.
[in]n[int] n specifies the number of columns of B. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A is not referenced and B need not be set before entry.
[in]APDevice pointer to matrix A on the GPU. A has dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPDevice pointer to the first matrix B_0 on the GPU. On entry, the leading m by n part of the array B must contain the matrix B, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m ).

◆ hipblasZtrmmBatched()

hipblasStatus_t hipblasZtrmmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
hipblasDoubleComplex *const  BP[],
int  ldb,
int  batchCount 
)

BLAS Level 3 API.

The hipBLAS trmm_batched API is from Legacy BLAS and it supports only in-place functionality. It is deprecated and it will be replaced with an API that supports both in-place and out-of-place functionality. The new API is available in hipBLAS versions 1.x.x and later. To get the new API compile with the directive -DHIPBLAS_V1.

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasStrmmBatched(hipblasHandle_t    handle,
                                        hipblasSideMode_t  side,
                                        hipblasFillMode_t  uplo,
                                        hipblasOperation_t transA,
                                        hipblasDiagType_t  diag,
                                        int                m,
                                        int                n,
                                        const float*       alpha,
                                        const float* const AP[],
                                        int                lda,
                                        float* const       BP[],
                                        int                ldb,
                                        int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasStrmmBatchedOutofplace(
                                        hipblasHandle_t    handle,
                                        hipblasSideMode_t  side,
                                        hipblasFillMode_t  uplo,
                                        hipblasOperation_t transA,
                                        hipblasDiagType_t  diag,
                                        int                m,
                                        int                n,
                                        const float*       alpha,
                                        const float* const AP[],
                                        int                lda,
                                        const float* const BP[],
                                        int                ldb,
                                        float* const       CP[],
                                        int                ldc,
                                        int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasDtrmmBatched(hipblasHandle_t     handle,
                                        hipblasSideMode_t   side,
                                        hipblasFillMode_t   uplo,
                                        hipblasOperation_t  transA,
                                        hipblasDiagType_t   diag,
                                        int                 m,
                                        int                 n,
                                        const double*       alpha,
                                        const double* const AP[],
                                        int                 lda,
                                        double* const       BP[],
                                        int                 ldb,
                                        int                 batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasDtrmmBatchedOutofplace(
                                        hipblasHandle_t     handle,
                                        hipblasSideMode_t   side,
                                        hipblasFillMode_t   uplo,
                                        hipblasOperation_t  transA,
                                        hipblasDiagType_t   diag,
                                        int                 m,
                                        int                 n,
                                        const double*       alpha,
                                        const double* const AP[],
                                        int                 lda,
                                        const double* const BP[],
                                        int                 ldb,
                                        double* const       CP[],
                                        int                 ldc,
                                        int                 batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasCtrmmBatched(hipblasHandle_t             handle,
                                        hipblasSideMode_t           side,
                                        hipblasFillMode_t           uplo,
                                        hipblasOperation_t          transA,
                                        hipblasDiagType_t           diag,
                                        int                         m,
                                        int                         n,
                                        const hipblasComplex*       alpha,
                                        const hipblasComplex* const AP[],
                                        int                         lda,
                                        hipblasComplex* const       BP[],
                                        int                         ldb,
                                        int                         batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasCtrmmBatchedOutofplace(
                                        hipblasHandle_t             handle,
                                        hipblasSideMode_t           side,
                                        hipblasFillMode_t           uplo,
                                        hipblasOperation_t          transA,
                                        hipblasDiagType_t           diag,
                                        int                         m,
                                        int                         n,
                                        const hipblasComplex*       alpha,
                                        const hipblasComplex* const AP[],
                                        int                         lda,
                                        const hipblasComplex* const BP[],
                                        int                         ldb,
                                        hipblasComplex* const       CP[],
                                        int                         ldc,
                                        int batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasZtrmmBatched(hipblasHandle_t                   handle,
                                        hipblasSideMode_t                 side,
                                        hipblasFillMode_t                 uplo,
                                        hipblasOperation_t                transA,
                                        hipblasDiagType_t                 diag,
                                        int                               m,
                                        int                               n,
                                        const hipblasDoubleComplex*       alpha,
                                        const hipblasDoubleComplex* const AP[],
                                        int                               lda,
                                        hipblasDoubleComplex* const       BP[],
                                        int                               ldb,
                                        int                               batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasZtrmmBatchedOutofplace(
                                        hipblasHandle_t                   handle,
                                        hipblasSideMode_t                 side,
                                        hipblasFillMode_t                 uplo,
                                        hipblasOperation_t                transA,
                                        hipblasDiagType_t                 diag,
                                        int                               m,
                                        int                               n,
                                        const hipblasDoubleComplex*       alpha,
                                        const hipblasDoubleComplex* const AP[],
                                        int                               lda,
                                        const hipblasDoubleComplex* const BP[],
                                        int                               ldb,
                                        hipblasDoubleComplex* const       CP[],
                                        int                               ldc,
                                        int                               batchCount);

#endif

The deprecated Legacy BLAS in-place trmm_batched performs one of the batched matrix-matrix operations:

B_i := alpha*op( A_i )*B_i,   or
B_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The new trmm_batched performs one of the matrix-matrix operations:

C_i := alpha*op( A_i )*B_i,   or
C_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The in-place functionality is still available in the new trmmm_batched by setting pointer C equal to pointer B and ldc equal to ldb.

alpha  is a scalar,  B_i  is an m by n matrix, C_i  is an m by n matrix,  A_i  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A_i )  is one  of

op( A_i ) = A_i   or   op( A_i ) = A_i^T   or   op( A_i ) = A_i^H.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A_i) multiplies B_i from the left or right as follows: HIPBLAS_SIDE_LEFT: B_i := alpha*op( A_i )*B_i. HIPBLAS_SIDE_RIGHT: B_i := alpha*B_i*op( A_i ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A_i) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A_i) = A_i. HIPBLAS_OP_T: op(A_i) = A_i^T. HIPBLAS_OP_C: op(A_i) = A_i^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A_i is unit triangular as follows: HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i is not referenced and B_i need not be set before entry.
[in]APDevice array of device pointers storing each matrix A_i on the GPU. Each A_i is of dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A_i are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice array of device pointers storing each matrix B_i on the GPU. On entry, the leading m by n part of the array B_i must contain the matrix B_i, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m ).
[in]batchCount[int] number of instances i in the batch.

◆ hipblasZtrmmBatchedOutofplace()

hipblasStatus_t hipblasZtrmmBatchedOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
const hipblasDoubleComplex *const  BP[],
int  ldb,
hipblasDoubleComplex *const  CP[],
int  ldc,
int  batchCount 
)

◆ hipblasZtrmmOutofplace()

hipblasStatus_t hipblasZtrmmOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
const hipblasDoubleComplex BP,
int  ldb,
hipblasDoubleComplex CP,
int  ldc 
)

◆ hipblasZtrmmStridedBatched()

hipblasStatus_t hipblasZtrmmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
hipblasDoubleComplex BP,
int  ldb,
hipblasStride  strideB,
int  batchCount 
)

BLAS Level 3 API.

The hipBLAS trmm_strided_batched API is from Legacy BLAS and it supports only in-place functionality. It is deprecated and it will be replaced with an API that supports both in-place and out-of-place functionality. The new API is available in hipBLAS versions 1.x.x and later. To get the new API compile with the directive -DHIPBLAS_V1.

#ifndef HIPBLAS_V1  //  deprecated
       hipblasStatus_t hipblasStrmmStridedBatched(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const float*       alpha,
                           const float*       AP,
                           int                lda,
                           hipblasStride      strideA,
                           float*             BP,
                           int                ldb,
                           hipblasStride      strideB,
                           int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasStrmmStridedBatchedOutofplace(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const float*       alpha,
                           const float*       AP,
                           int                lda,
                           hipblasStride      strideA,
                           const float*       BP,
                           int                ldb,
                           hipblasStride      strideB,
                           float*             CP,
                           int                ldc,
                           hipblasStride      strideC,
                           int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated
       hipblasStatus_t hipblasDtrmmStridedBatched(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const double*      alpha,
                           const double*      AP,
                           int                lda,
                           hipblasStride      strideA,
                           double*            BP,
                           int                ldb,
                           hipblasStride      strideB,
                           int                batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasDtrmmStridedBatchedOutofplace(
                           hipblasHandle_t    handle,
                           hipblasSideMode_t  side,
                           hipblasFillMode_t  uplo,
                           hipblasOperation_t transA,
                           hipblasDiagType_t  diag,
                           int                m,
                           int                n,
                           const double*      alpha,
                           const double*      AP,
                           int                lda,
                           hipblasStride      strideA,
                           const double*      BP,
                           int                ldb,
                           hipblasStride      strideB,
                           double*            CP,
                           int                ldc,
                           hipblasStride      strideC,
                           int                batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

       hipblasStatus_t hipblasCtrmmStridedBatched(
                           hipblasHandle_t       handle,
                           hipblasSideMode_t     side,
                           hipblasFillMode_t     uplo,
                           hipblasOperation_t    transA,
                           hipblasDiagType_t     diag,
                           int                   m,
                           int                   n,
                           const hipblasComplex* alpha,
                           const hipblasComplex* AP,
                           int                   lda,
                           hipblasStride         strideA,
                           hipblasComplex*       BP,
                           int                   ldb,
                           hipblasStride         strideB,
                           int                   batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

       hipblasStatus_t hipblasCtrmmStridedBatchedOutofplace(
                           hipblasHandle_t       handle,
                           hipblasSideMode_t     side,
                           hipblasFillMode_t     uplo,
                           hipblasOperation_t    transA,
                           hipblasDiagType_t     diag,
                           int                   m,
                           int                   n,
                           const hipblasComplex* alpha,
                           const hipblasComplex* AP,
                           int                   lda,
                           hipblasStride         strideA,
                           const hipblasComplex* BP,
                           int                   ldb,
                           hipblasStride         strideB,
                           hipblasComplex*       CP,
                           int                   ldc,
                           hipblasStride         strideC,
                           int batchCount);

#endif

#ifndef HIPBLAS_V1  //  deprecated

    hipblasStatus_t hipblasZtrmmStridedBatched(hipblasHandle_t             handle,
                                               hipblasSideMode_t           side,
                                               hipblasFillMode_t           uplo,
                                               hipblasOperation_t          transA,
                                               hipblasDiagType_t           diag,
                                               int                         m,
                                               int                         n,
                                               const hipblasDoubleComplex* alpha,
                                               const hipblasDoubleComplex* AP,
                                               int                         lda,
                                               hipblasStride               strideA,
                                               hipblasDoubleComplex*       BP,
                                               int                         ldb,
                                               hipblasStride               strideB,
                                               int                         batchCount);

#else  //  available in hipBLAS version 1.x.x and later with -DHIPBLAS_V1

    hipblasStatus_t hipblasZtrmmStridedBatchedOutofplace(
                                                hipblasHandle_t             handle,
                                                hipblasSideMode_t           side,
                                                hipblasFillMode_t           uplo,
                                                hipblasOperation_t          transA,
                                                hipblasDiagType_t           diag,
                                                int                         m,
                                                int                         n,
                                                const hipblasDoubleComplex* alpha,
                                                const hipblasDoubleComplex* AP,
                                                int                         lda,
                                                hipblasStride               strideA,
                                                const hipblasDoubleComplex* BP,
                                                int                         ldb,
                                                hipblasStride               strideB,
                                                hipblasDoubleComplex*       BC,
                                                int                         ldc,
                                                hipblasStride               strideC,
                                                int                         batchCount);

#endif

The deprecated Legacy BLAS in-place trmm_strided_batched performs one of the strided_batched matrix-matrix operations:

B_i := alpha*op( A_i )*B_i,   or
B_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The new trmm_batched performs one of the matrix-matrix operations:

C_i := alpha*op( A_i )*B_i,   or
C_i := alpha*B_i*op( A_i )  for i = 0, 1, ... batch_count -1,

The in-place functionality is still available in the new trmmm_batched by setting pointer C equal to pointer B, setting ldc equal to ldb, and setting stride_C equal to stride_B.

alpha  is a scalar,  B_i  is an m by n matrix, C_i  is an m by n matrix,  A_i  is a unit, or
non-unit,  upper or lower triangular matrix  and  op( A_i )  is one  of

op( A_i ) = A_i   or
op( A_i ) = A_i^T   or
op( A_i ) = A_i^H.
  • Supported precisions in hipBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] Specifies whether op(A_i) multiplies B_i from the left or right as follows: HIPBLAS_SIDE_LEFT: B_i := alpha*op( A_i )*B_i. HIPBLAS_SIDE_RIGHT: B_i := alpha*B_i*op( A_i ).
[in]uplo[hipblasFillMode_t] Specifies whether the matrix A is an upper or lower triangular matrix as follows: HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] Specifies the form of op(A_i) to be used in the matrix multiplication as follows: HIPBLAS_OP_N: op(A_i) = A_i. HIPBLAS_OP_T: op(A_i) = A_i^T. HIPBLAS_OP_C: op(A_i) = A_i^H.
[in]diag[hipblasDiagType_t] Specifies whether or not A_i is unit triangular as follows: HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B_i. m >= 0.
[in]n[int] n specifies the number of columns of B_i. n >= 0.
[in]alphaalpha specifies the scalar alpha. When alpha is zero then A_i is not referenced and B_i need not be set before entry.
[in]APDevice pointer to the first matrix A_0 on the GPU. Each A_i is of dimension ( lda, k ), where k is m when side == HIPBLAS_SIDE_LEFT and is n when side == HIPBLAS_SIDE_RIGHT.

When uplo == HIPBLAS_FILL_MODE_UPPER the leading k by k upper triangular part of the array A must contain the upper triangular matrix and the strictly lower triangular part of A is not referenced.

When uplo == HIPBLAS_FILL_MODE_LOWER the leading k by k lower triangular part of the array A must contain the lower triangular matrix and the strictly upper triangular part of A is not referenced.

Note that when diag == HIPBLAS_DIAG_UNIT the diagonal elements of A_i are not referenced either, but are assumed to be unity.

Parameters
[in]lda[int] lda specifies the first dimension of A. if side == HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side == HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one matrix (A_i) and the next one (A_i+1)
[in,out]BPDevice pointer to the first matrix B_0 on the GPU. On entry, the leading m by n part of the array B_i must contain the matrix B_i, and on exit is overwritten by the transformed matrix.
[in]ldb[int] ldb specifies the first dimension of B_i. ldb >= max( 1, m ).
[in]strideB[hipblasStride] stride from the start of one matrix (B_i) and the next one (B_i+1)
[in]batchCount[int] number of instances i in the batch.

◆ hipblasZtrmmStridedBatchedOutofplace()

hipblasStatus_t hipblasZtrmmStridedBatchedOutofplace ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
const hipblasDoubleComplex BP,
int  ldb,
hipblasStride  strideB,
hipblasDoubleComplex BC,
int  ldc,
hipblasStride  strideC,
int  batchCount 
)

◆ hipblasZtrmv()

hipblasStatus_t hipblasZtrmv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex AP,
int  lda,
hipblasDoubleComplex x,
int  incx 
)

BLAS Level 2 API.

trmv performs one of the matrix-vector operations

 x = A*x or x = A**T*x,

where x is an n element vector and A is an n by n unit, or non-unit, upper or lower triangular matrix.

The vector x is overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of A. m >= 0.
[in]APdevice pointer storing matrix A, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A. lda = max( 1, m ).
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasZtrmvBatched()

hipblasStatus_t hipblasZtrmvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex *const  AP[],
int  lda,
hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

trmvBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix)

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer storing pointer of matrices A_i, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A_i. lda >= max( 1, m ).
[in]xdevice pointer storing vectors x_i.
[in]incx[int] specifies the increment for the elements of vectors x_i.
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasZtrmvStridedBatched()

hipblasStatus_t hipblasZtrmvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

trmvStridedBatched performs one of the matrix-vector operations

 x_i = A_i*x_i or x_i = A**T*x_i, 0 \le i < batchCount

where x_i is an n element vector and A_i is an n by n (unit, or non-unit, upper or lower triangular matrix) with strides specifying how to retrieve $x_i$ (resp. $A_i$) from $x_{i-1}$ (resp. $A_i$).

The vectors x_i are overwritten.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of matrices A_i. m >= 0.
[in]APdevice pointer of the matrix A_0, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A_i. lda >= max( 1, m ).
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_{i + 1}
[in]xdevice pointer storing the vector x_0.
[in]incx[int] specifies the increment for the elements of one vector x.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_{i + 1}
[in]batchCount[int] The number of batched matrices/vectors.

◆ hipblasZtrsm()

hipblasStatus_t hipblasZtrsm ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasDoubleComplex BP,
int  ldb 
)

BLAS Level 3 API.

trsm solves

op(A)*X = alpha*B or  X*op(A) = alpha*B,

where alpha is a scalar, X and B are m by n matrices, A is triangular matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

The matrix X is overwritten on B.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time.

(where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of B. m >= 0.
[in]n[int] n specifies the number of columns of B. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice pointer storing matrix A. of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of A. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice pointer storing matrix B.
[in]ldb[int] ldb specifies the first dimension of B. ldb >= max( 1, m ).

◆ hipblasZtrsmBatched()

hipblasStatus_t hipblasZtrsmBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex *const  AP[],
int  lda,
hipblasDoubleComplex *const  BP[],
int  ldb,
int  batchCount 
)

BLAS Level 3 API.

trsmBatched performs the following batched operation:

op(A_i)*X_i = alpha*B_i or  X_i*op(A_i) = alpha*B_i, for i = 1, ..., batchCount.

where alpha is a scalar, X and B are batched m by n matrices, A is triangular batched matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

Each matrix X_i is overwritten on B_i for i = 1, ..., batchCount.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time. (where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each B_i. m >= 0.
[in]n[int] n specifies the number of columns of each B_i. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice array of device pointers storing each matrix A_i on the GPU. Matricies are of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of each A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in,out]BPdevice array of device pointers storing each matrix B_i on the GPU.
[in]ldb[int] ldb specifies the first dimension of each B_i. ldb >= max( 1, m ).
[in]batchCount[int] number of trsm operatons in the batch.

◆ hipblasZtrsmStridedBatched()

hipblasStatus_t hipblasZtrsmStridedBatched ( hipblasHandle_t  handle,
hipblasSideMode_t  side,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
int  n,
const hipblasDoubleComplex alpha,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
hipblasDoubleComplex BP,
int  ldb,
hipblasStride  strideB,
int  batchCount 
)

BLAS Level 3 API.

trsmSridedBatched performs the following strided batched operation:

op(A_i)*X_i = alpha*B_i or  X_i*op(A_i) = alpha*B_i, for i = 1, ..., batchCount.

where alpha is a scalar, X and B are strided batched m by n matrices, A is triangular strided batched matrix and op(A) is one of

op( A ) = A   or   op( A ) = A^T   or   op( A ) = A^H.

Each matrix X_i is overwritten on B_i for i = 1, ..., batchCount.

Note about memory allocation: When trsm is launched with a k evenly divisible by the internal block size of 128, and is no larger than 10 of these blocks, the API takes advantage of utilizing pre-allocated memory found in the handle to increase overall performance. This memory can be managed by using the environment variable WORKBUF_TRSM_B_CHNK. When this variable is not set the device memory used for temporary storage will default to 1 MB and may result in chunking, which in turn may reduce performance. Under these circumstances it is recommended that WORKBUF_TRSM_B_CHNK be set to the desired chunk of right hand sides to be used at a time. (where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT)

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]side[hipblasSideMode_t] HIPBLAS_SIDE_LEFT: op(A)*X = alpha*B. HIPBLAS_SIDE_RIGHT: X*op(A) = alpha*B.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: each A_i is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: each A_i is a lower triangular matrix.
[in]transA[hipblasOperation_t] HIPBLAS_OP_N: op(A) = A. HIPBLAS_OP_T: op(A) = A^T. HIPBLAS_OP_C: op(A) = A^H.
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: each A_i is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: each A_i is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each B_i. m >= 0.
[in]n[int] n specifies the number of columns of each B_i. n >= 0.
[in]alphadevice pointer or host pointer specifying the scalar alpha. When alpha is &zero then A is not referenced and B need not be set before entry.
[in]APdevice pointer pointing to the first matrix A_1. of dimension ( lda, k ), where k is m when HIPBLAS_SIDE_LEFT and is n when HIPBLAS_SIDE_RIGHT only the upper/lower triangular part is accessed.
[in]lda[int] lda specifies the first dimension of each A_i. if side = HIPBLAS_SIDE_LEFT, lda >= max( 1, m ), if side = HIPBLAS_SIDE_RIGHT, lda >= max( 1, n ).
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1).
[in,out]BPdevice pointer pointing to the first matrix B_1.
[in]ldb[int] ldb specifies the first dimension of each B_i. ldb >= max( 1, m ).
[in]strideB[hipblasStride] stride from the start of one B_i matrix to the next B_(i + 1).
[in]batchCount[int] number of trsm operatons in the batch.

◆ hipblasZtrsv()

hipblasStatus_t hipblasZtrsv ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex AP,
int  lda,
hipblasDoubleComplex x,
int  incx 
)

BLAS Level 2 API.

trsv solves

 A*x = b or A**T*x = b,

where x and b are vectors and A is a triangular matrix.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : s,d,c,z
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice pointer storing matrix A, of dimension ( lda, m )
[in]lda[int] specifies the leading dimension of A. lda = max( 1, m ).
[in]xdevice pointer storing vector x.
[in]incx[int] specifies the increment for the elements of x.

◆ hipblasZtrsvBatched()

hipblasStatus_t hipblasZtrsvBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex *const  AP[],
int  lda,
hipblasDoubleComplex *const  x[],
int  incx,
int  batchCount 
)

BLAS Level 2 API.

trsvBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i,

where (A_i, x_i, b_i) is the i-th instance of the batch. x_i and b_i are vectors and A_i is an m by m triangular matrix.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of b. m >= 0.
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i. lda = max(1, m)
[in]xdevice array of device pointers storing each vector x_i.
[in]incx[int] specifies the increment for the elements of x.
[in]batchCount[int] number of instances in the batch

◆ hipblasZtrsvStridedBatched()

hipblasStatus_t hipblasZtrsvStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasOperation_t  transA,
hipblasDiagType_t  diag,
int  m,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
hipblasDoubleComplex x,
int  incx,
hipblasStride  stridex,
int  batchCount 
)

BLAS Level 2 API.

trsvStridedBatched solves

 A_i*x_i = b_i or A_i**T*x_i = b_i,

where (A_i, x_i, b_i) is the i-th instance of the batch. x_i and b_i are vectors and A_i is an m by m triangular matrix, for i = 1, ..., batchCount.

The vector x is overwritten on b.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] HIPBLAS_FILL_MODE_UPPER: A is an upper triangular matrix. HIPBLAS_FILL_MODE_LOWER: A is a lower triangular matrix.
[in]transA[hipblasOperation_t]
[in]diag[hipblasDiagType_t] HIPBLAS_DIAG_UNIT: A is assumed to be unit triangular. HIPBLAS_DIAG_NON_UNIT: A is not assumed to be unit triangular.
[in]m[int] m specifies the number of rows of each b_i. m >= 0.
[in]APdevice pointer to the first matrix (A_1) in the batch, of dimension ( lda, m )
[in]strideA[hipblasStride] stride from the start of one A_i matrix to the next A_(i + 1)
[in]lda[int] specifies the leading dimension of each A_i. lda = max( 1, m ).
[in,out]xdevice pointer to the first vector (x_1) in the batch.
[in]stridex[hipblasStride] stride from the start of one x_i vector to the next x_(i + 1)
[in]incx[int] specifies the increment for the elements of each x_i.
[in]batchCount[int] number of instances in the batch

◆ hipblasZtrtri()

hipblasStatus_t hipblasZtrtri ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const hipblasDoubleComplex AP,
int  lda,
hipblasDoubleComplex invA,
int  ldinvA 
)

BLAS Level 3 API.

trtri compute the inverse of a matrix A, namely, invA

and write the result into invA;
  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER' if HIPBLAS_FILL_MODE_UPPER, the lower part of A is not referenced if HIPBLAS_FILL_MODE_LOWER, the upper part of A is not referenced
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int] size of matrix A and invA
[in]APdevice pointer storing matrix A.
[in]lda[int] specifies the leading dimension of A.
[out]invAdevice pointer storing matrix invA.
[in]ldinvA[int] specifies the leading dimension of invA.

◆ hipblasZtrtriBatched()

hipblasStatus_t hipblasZtrtriBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const hipblasDoubleComplex *const  AP[],
int  lda,
hipblasDoubleComplex invA[],
int  ldinvA,
int  batchCount 
)

BLAS Level 3 API.

trtriBatched compute the inverse of A_i and write into invA_i where A_i and invA_i are the i-th matrices in the batch, for i = 1, ..., batchCount.

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER'
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int]
[in]APdevice array of device pointers storing each matrix A_i.
[in]lda[int] specifies the leading dimension of each A_i.
[out]invAdevice array of device pointers storing the inverse of each matrix A_i. Partial inplace operation is supported, see below. If UPLO = 'U', the leading N-by-N upper triangular part of the invA will store the inverse of the upper triangular matrix, and the strictly lower triangular part of invA is cleared. If UPLO = 'L', the leading N-by-N lower triangular part of the invA will store the inverse of the lower triangular matrix, and the strictly upper triangular part of invA is cleared.
[in]ldinvA[int] specifies the leading dimension of each invA_i.
[in]batchCount[int] numbers of matrices in the batch

◆ hipblasZtrtriStridedBatched()

hipblasStatus_t hipblasZtrtriStridedBatched ( hipblasHandle_t  handle,
hipblasFillMode_t  uplo,
hipblasDiagType_t  diag,
int  n,
const hipblasDoubleComplex AP,
int  lda,
hipblasStride  strideA,
hipblasDoubleComplex invA,
int  ldinvA,
hipblasStride  stride_invA,
int  batchCount 
)

BLAS Level 3 API.

trtriStridedBatched compute the inverse of A_i and write into invA_i where A_i and invA_i are the i-th matrices in the batch, for i = 1, ..., batchCount

  • Supported precisions in rocBLAS : s,d,c,z
  • Supported precisions in cuBLAS : No support
Parameters
[in]handle[hipblasHandle_t] handle to the hipblas library context queue.
[in]uplo[hipblasFillMode_t] specifies whether the upper 'HIPBLAS_FILL_MODE_UPPER' or lower 'HIPBLAS_FILL_MODE_LOWER'
[in]diag[hipblasDiagType_t] = 'HIPBLAS_DIAG_NON_UNIT', A is non-unit triangular; = 'HIPBLAS_DIAG_UNIT', A is unit triangular;
[in]n[int]
[in]APdevice pointer pointing to address of first matrix A_1.
[in]lda[int] specifies the leading dimension of each A.
[in]strideA[hipblasStride] "batch stride a": stride from the start of one A_i matrix to the next A_(i + 1).
[out]invAdevice pointer storing the inverses of each matrix A_i. Partial inplace operation is supported, see below. If UPLO = 'U', the leading N-by-N upper triangular part of the invA will store the inverse of the upper triangular matrix, and the strictly lower triangular part of invA is cleared. If UPLO = 'L', the leading N-by-N lower triangular part of the invA will store the inverse of the lower triangular matrix, and the strictly upper triangular part of invA is cleared.
[in]ldinvA[int] specifies the leading dimension of each invA_i.
[in]stride_invA[hipblasStride] "batch stride invA": stride from the start of one invA_i matrix to the next invA_(i + 1).
[in]batchCount[int] numbers of matrices in the batch